Hello everyone!
\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I downloaded the articles that seemed relevant after reading their summaries.
\nPapers:
\nInstance Segmentation of Neural Cells - Instance segmentation of neural cells plays an important role in brain study. However, this task is challenging due to the special shapes and behaviors of neural cells. Existing methods are not precise enough to capture their tiny structures, e.g., filopodia and lamellipodia, which are critical to the understanding of cell interaction and behavior. To this end, we propose a novel deep multi-task learning model to jointly detect and segment neural cells instance-wise. Our method is built upon SSD, with ResNet101 as the backbone to achieve both high detection accuracy and fast speed. Furthermore, unlike existing works which tend to produce wavy and inaccurate boundaries, we embed a deconvolution module into SSD to better capture details. Experiments on a dataset of neural cell microscopic images show that our method is able to achieve better performance in terms of accuracy and efficiency, comparing favorably with current state-of-the-art methods.
Mask-RCNN for Cell Instance Segmentation - We proposed an automatic nucleus segmentation algorithm of H&E stained tissue microscopy images. Mask-RCNN is a recently proposed state-of-the-art algorithm for object detection and object instance segmentation of natural images. In this paper, we demonstrate that Mask-RCNN can be used to perform highly effective and efficient automatic segmentation of H&E microscopy images for cell nuclei. We propose a novel MASK Non-maximum suppression (NMS) module which can automatically ensemble classifiers results and increase the robustness of model.
IRNet: Instance Relation Network for\nOverlapping Cervical Cell Segmentation - Cell instance segmentation in Pap smear image remains challenging due to the wide existence of occlusion among translucent cytoplasm in cell clumps. Conventional methods heavily rely on accurate nuclei detection results and are easily disturbed by miscellaneous objects. In this paper, we propose a novel Instance Relation Network (IRNet) for robust overlapping cell segmentation by exploring instance relation interaction. Specifically, we propose the Instance Relation Module to construct the cell association matrix for transferring information among individual cell-instance features. With the collaboration of different instances, the augmented features gain benefits from contextual information and improve semantic consistency. Meanwhile, we proposed a sparsity constrained Duplicate Removal Module to eliminate the misalignment between classification and localization accuracy for candidates selection. The largest cervical Pap smear (CPS) dataset with more than 8000 cell annotations in Pap smear image was constructed for comprehensive evaluation. Our method outperforms other methods by a large margin, demonstrating the effectiveness of exploring instance relation.
Weakly Supervised Cell Instance Segmentation by Propagating from Detection Response - Cell shape analysis is important in biomedical research. Deep learning methods may perform to segment individual cells if they use sufficient training data that the boundary of each cell is annotated. However, it is very time-consuming for preparing such detailed annotation for many cell culture conditions. In this paper, we propose a weakly supervised method that can segment individual cell regions who touch each other with unclear boundaries in dense conditions without the training data for cell regions. We demonstrated the efficacy of our method using several data-set including multiple cell types captured by several types of microscopy. Our method achieved the highest accuracy compared with several conventional methods. In addition, we demonstrated that our method can perform without any annotation by using fluorescence images that cell nuclear were stained as training data.
Deep Semi-supervised Knowledge Distillation for\nOverlapping Cervical Cell Instance Segmentation - Deep learning methods show promising results for overlapping cervical cell instance segmentation. However, in order to train a model with good generalization ability, voluminous pixel-level annotations are demanded which is quite expensive and time-consuming for acquisition. In this paper, we propose to leverage both labeled and unlabeled data for instance segmentation with improved accuracy by knowledge distillation. We propose a novel Mask-guided Mean Teacher framework with Perturbation-sensitive Sample Mining (MMT-PSM), which consists of a teacher and a student network during training. Two networks are encouraged to be consistent both in feature and semantic level under small perturbations. The teacherโs self-ensemble predictions from K-time augmented samples are used to construct the reliable pseudolabels for optimizing the student. We design a novel strategy to estimate the sensitivity to perturbations for each proposal and select informative samples from massive cases to facilitate fast and effective semantic distillation. In addition, to eliminate the unavoidable noise from the background region, we propose to use the predicted segmentation mask as guidance to enforce the feature distillation in the foreground region. Experiments show that the proposed method improves the performance significantly compared with the supervised method learned from labeled data only, and outperforms state-of-the-art semi-supervised methods.
Attentive neural cell instance segmentation - Neural cell instance segmentation, which aims at joint detection and segmentation of every neural cell in a microscopic image, is essential to many neuroscience applications. The challenge of this task in- volves cell adhesion, cell distortion, unclear cell contours, low-contrast cell protrusion structures, and background impurities. Consequently, current instance segmentation methods generally fall short of pre- cision. In this paper, we propose an attentive instance segmentation method that accurately predicts the bounding box of each cell as well as its segmentation mask simultaneously. In particular, our method builds on a joint network that combines a single shot multi-box detector (SSD) and a U-net. Furthermore, we employ the attention mechanism in both detection and segmentation modules to focus the model on the useful features. The proposed method is validated on a dataset of neural cell microscopic images. Experimental results demonstrate that our approach can accurately detect and segment neural cell in- stances at a fast speed, comparing favorably with the state-of-the-art methods.
LIVECellโA large-scale dataset for label-free live cell segmentation - Light microscopy combined with well-established protocols of two-dimensional cell culture facilitates high-throughput quantitative imaging to study biological phenomena. Accurate segmentation of individual cells in images enables exploration of complex biological questions, but can require sophisticated imaging processing pipelines in cases of low contrast and high object density. Deep learning-based methods are considered state-of-the-art for image segmentation but typically require vast amounts of annotated data, for which there is no suitable resource available in the field of label-free cellular imaging. Here, we present LIVECell, a large, high-quality, manually annotated and expert-validated dataset of phase-contrast images, consisting of over 1.6 million cells from a diverse set of cell morphologies and culture densities. To further demonstrate its use, we train convolutional neural network-based models using LIVECell and evaluate model segmentation accuracy with a proposed a suite of benchmarks. Thanks to @maxwell110
Nuclei Segmentation via a Deep Panoptic Model with Semantic Feature Fusion - Automated detection and segmentation of individual nuclei in histopathology images is important for cancer diagnosis and prognosis. Due to the high variability of nuclei appearances and numerous overlapping objects, this task still remains challenging. Deep learning based semantic and instance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent information loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin. Thanks to @osamurai
Pixel-wise neural cell instance segmentation - Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin. Thanks to @abhishandy
๐ U-net with Pytorch
\n๐ Using U-net for a Kaggle competition (AutoEncoder)
๐ Model trained by LIVECell for transfer learning Thanks to @maxwell110
\n๐ FIAR's Dectron 2 with Pytorch Thanks to @maxwell110
Have a good competition and don't hesitate to comment!
", "messages": [ { "id": "1545410", "postDate": "10/15/2021 07:39:57", "content": "Hello everyone!
\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I downloaded the articles that seemed relevant after reading their summaries.
\nPapers:
\nInstance Segmentation of Neural Cells - Instance segmentation of neural cells plays an important role in brain study. However, this task is challenging due to the special shapes and behaviors of neural cells. Existing methods are not precise enough to capture their tiny structures, e.g., filopodia and lamellipodia, which are critical to the understanding of cell interaction and behavior. To this end, we propose a novel deep multi-task learning model to jointly detect and segment neural cells instance-wise. Our method is built upon SSD, with ResNet101 as the backbone to achieve both high detection accuracy and fast speed. Furthermore, unlike existing works which tend to produce wavy and inaccurate boundaries, we embed a deconvolution module into SSD to better capture details. Experiments on a dataset of neural cell microscopic images show that our method is able to achieve better performance in terms of accuracy and efficiency, comparing favorably with current state-of-the-art methods.
Mask-RCNN for Cell Instance Segmentation - We proposed an automatic nucleus segmentation algorithm of H&E stained tissue microscopy images. Mask-RCNN is a recently proposed state-of-the-art algorithm for object detection and object instance segmentation of natural images. In this paper, we demonstrate that Mask-RCNN can be used to perform highly effective and efficient automatic segmentation of H&E microscopy images for cell nuclei. We propose a novel MASK Non-maximum suppression (NMS) module which can automatically ensemble classifiers results and increase the robustness of model.
IRNet: Instance Relation Network for\nOverlapping Cervical Cell Segmentation - Cell instance segmentation in Pap smear image remains challenging due to the wide existence of occlusion among translucent cytoplasm in cell clumps. Conventional methods heavily rely on accurate nuclei detection results and are easily disturbed by miscellaneous objects. In this paper, we propose a novel Instance Relation Network (IRNet) for robust overlapping cell segmentation by exploring instance relation interaction. Specifically, we propose the Instance Relation Module to construct the cell association matrix for transferring information among individual cell-instance features. With the collaboration of different instances, the augmented features gain benefits from contextual information and improve semantic consistency. Meanwhile, we proposed a sparsity constrained Duplicate Removal Module to eliminate the misalignment between classification and localization accuracy for candidates selection. The largest cervical Pap smear (CPS) dataset with more than 8000 cell annotations in Pap smear image was constructed for comprehensive evaluation. Our method outperforms other methods by a large margin, demonstrating the effectiveness of exploring instance relation.
Weakly Supervised Cell Instance Segmentation by Propagating from Detection Response - Cell shape analysis is important in biomedical research. Deep learning methods may perform to segment individual cells if they use sufficient training data that the boundary of each cell is annotated. However, it is very time-consuming for preparing such detailed annotation for many cell culture conditions. In this paper, we propose a weakly supervised method that can segment individual cell regions who touch each other with unclear boundaries in dense conditions without the training data for cell regions. We demonstrated the efficacy of our method using several data-set including multiple cell types captured by several types of microscopy. Our method achieved the highest accuracy compared with several conventional methods. In addition, we demonstrated that our method can perform without any annotation by using fluorescence images that cell nuclear were stained as training data.
Deep Semi-supervised Knowledge Distillation for\nOverlapping Cervical Cell Instance Segmentation - Deep learning methods show promising results for overlapping cervical cell instance segmentation. However, in order to train a model with good generalization ability, voluminous pixel-level annotations are demanded which is quite expensive and time-consuming for acquisition. In this paper, we propose to leverage both labeled and unlabeled data for instance segmentation with improved accuracy by knowledge distillation. We propose a novel Mask-guided Mean Teacher framework with Perturbation-sensitive Sample Mining (MMT-PSM), which consists of a teacher and a student network during training. Two networks are encouraged to be consistent both in feature and semantic level under small perturbations. The teacherโs self-ensemble predictions from K-time augmented samples are used to construct the reliable pseudolabels for optimizing the student. We design a novel strategy to estimate the sensitivity to perturbations for each proposal and select informative samples from massive cases to facilitate fast and effective semantic distillation. In addition, to eliminate the unavoidable noise from the background region, we propose to use the predicted segmentation mask as guidance to enforce the feature distillation in the foreground region. Experiments show that the proposed method improves the performance significantly compared with the supervised method learned from labeled data only, and outperforms state-of-the-art semi-supervised methods.
Attentive neural cell instance segmentation - Neural cell instance segmentation, which aims at joint detection and segmentation of every neural cell in a microscopic image, is essential to many neuroscience applications. The challenge of this task in- volves cell adhesion, cell distortion, unclear cell contours, low-contrast cell protrusion structures, and background impurities. Consequently, current instance segmentation methods generally fall short of pre- cision. In this paper, we propose an attentive instance segmentation method that accurately predicts the bounding box of each cell as well as its segmentation mask simultaneously. In particular, our method builds on a joint network that combines a single shot multi-box detector (SSD) and a U-net. Furthermore, we employ the attention mechanism in both detection and segmentation modules to focus the model on the useful features. The proposed method is validated on a dataset of neural cell microscopic images. Experimental results demonstrate that our approach can accurately detect and segment neural cell in- stances at a fast speed, comparing favorably with the state-of-the-art methods.
LIVECellโA large-scale dataset for label-free live cell segmentation - Light microscopy combined with well-established protocols of two-dimensional cell culture facilitates high-throughput quantitative imaging to study biological phenomena. Accurate segmentation of individual cells in images enables exploration of complex biological questions, but can require sophisticated imaging processing pipelines in cases of low contrast and high object density. Deep learning-based methods are considered state-of-the-art for image segmentation but typically require vast amounts of annotated data, for which there is no suitable resource available in the field of label-free cellular imaging. Here, we present LIVECell, a large, high-quality, manually annotated and expert-validated dataset of phase-contrast images, consisting of over 1.6 million cells from a diverse set of cell morphologies and culture densities. To further demonstrate its use, we train convolutional neural network-based models using LIVECell and evaluate model segmentation accuracy with a proposed a suite of benchmarks. Thanks to @maxwell110
Nuclei Segmentation via a Deep Panoptic Model with Semantic Feature Fusion - Automated detection and segmentation of individual nuclei in histopathology images is important for cancer diagnosis and prognosis. Due to the high variability of nuclei appearances and numerous overlapping objects, this task still remains challenging. Deep learning based semantic and instance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent information loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin. Thanks to @osamurai
Pixel-wise neural cell instance segmentation - Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin. Thanks to @abhishandy
๐ U-net with Pytorch
\n๐ Using U-net for a Kaggle competition (AutoEncoder)
๐ Model trained by LIVECell for transfer learning Thanks to @maxwell110
\n๐ FIAR's Dectron 2 with Pytorch Thanks to @maxwell110
Have a good competition and don't hesitate to comment!
", "rawMarkdown": "Hello everyone!\n\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I downloaded the articles that seemed relevant after reading their summaries.\n\n**Papers:**\n\n- [Instance Segmentation of Neural Cells](https://www.researchgate.net/publication/329933921_Instance_Segmentation_of_Neural_Cells) - Instance segmentation of neural cells plays an important role in brain study. However, this task is challenging due to the special shapes and behaviors of neural cells. Existing methods are not precise enough to capture their tiny structures, e.g., filopodia and lamellipodia, which are critical to the understanding of cell interaction and behavior. To this end, we propose a novel deep multi-task learning model to jointly detect and segment neural cells instance-wise. Our method is built upon SSD, with ResNet101 as the backbone to achieve both high detection accuracy and fast speed. Furthermore, unlike existing works which tend to produce wavy and inaccurate boundaries, we embed a deconvolution module into SSD to better capture details. Experiments on a dataset of neural cell microscopic images show that our method is able to achieve better performance in terms of accuracy and efficiency, comparing favorably with current state-of-the-art methods.\n\n- [Mask-RCNN for Cell Instance Segmentation](https://www.researchgate.net/publication/340998089_Rank_6_Mask-RCNN_for_Cell_Instance_Segmentation) - We proposed an automatic nucleus segmentation algorithm of H&E stained tissue microscopy images. Mask-RCNN is a recently proposed state-of-the-art algorithm for object detection and object instance segmentation of natural images. In this paper, we demonstrate that Mask-RCNN can be used to perform highly effective and efficient automatic segmentation of H&E microscopy images for cell nuclei. We propose a novel MASK Non-maximum suppression (NMS) module which can automatically ensemble classifiers results and increase the robustness of model.\n\n- [IRNet: Instance Relation Network for\nOverlapping Cervical Cell Segmentation](https://arxiv.org/abs/1908.06623) - Cell instance segmentation in Pap smear image remains challenging due to the wide existence of occlusion among translucent cytoplasm in cell clumps. Conventional methods heavily rely on accurate nuclei detection results and are easily disturbed by miscellaneous objects. In this paper, we propose a novel Instance Relation Network (IRNet) for robust overlapping cell segmentation by exploring instance relation interaction. Specifically, we propose the Instance Relation Module to construct the cell association matrix for transferring information among individual cell-instance features. With the collaboration of different instances, the augmented features gain benefits from contextual information and improve semantic consistency. Meanwhile, we proposed a sparsity constrained Duplicate Removal Module to eliminate the misalignment between classification and localization accuracy for candidates selection. The largest cervical Pap smear (CPS) dataset with more than 8000 cell annotations in Pap smear image was constructed for comprehensive evaluation. Our method outperforms other methods by a large margin, demonstrating the effectiveness of exploring instance relation. \n\n- [Weakly Supervised Cell Instance Segmentation by Propagating from Detection Response](https://www.researchgate.net/publication/336392797_Weakly_Supervised_Cell_Instance_Segmentation_by_Propagating_from_Detection_Response) - Cell shape analysis is important in biomedical research. Deep learning methods may perform to segment individual cells if they use sufficient training data that the boundary of each cell is annotated. However, it is very time-consuming for preparing such detailed annotation for many cell culture conditions. In this paper, we propose a weakly supervised method that can segment individual cell regions who touch each other with unclear boundaries in dense conditions without the training data for cell regions. We demonstrated the efficacy of our method using several data-set including multiple cell types captured by several types of microscopy. Our method achieved the highest accuracy compared with several conventional methods. In addition, we demonstrated that our method can perform without any annotation by using fluorescence images that cell nuclear were stained as training data.\n\n- [Deep Semi-supervised Knowledge Distillation for\nOverlapping Cervical Cell Instance Segmentation](https://arxiv.org/abs/2007.10787) - Deep learning methods show promising results for overlapping cervical cell instance segmentation. However, in order to train a model with good generalization ability, voluminous pixel-level annotations are demanded which is quite expensive and time-consuming for acquisition. In this paper, we propose to leverage both labeled and unlabeled data for instance segmentation with improved accuracy by knowledge distillation. We propose a novel Mask-guided Mean Teacher framework with Perturbation-sensitive Sample Mining (MMT-PSM), which consists of a teacher and a student network during training. Two networks are encouraged to be consistent both in feature and semantic level under small perturbations. The teacherโs self-ensemble predictions from K-time augmented samples are used to construct the reliable pseudolabels for optimizing the student. We design a novel strategy to estimate the sensitivity to perturbations for each proposal and select informative samples from massive cases to facilitate fast and effective semantic distillation. In addition, to eliminate the unavoidable noise from the background region, we propose to use the predicted segmentation mask as guidance to enforce the feature distillation in the foreground region. Experiments show that the proposed method improves the performance significantly compared with the supervised method learned from labeled data only, and outperforms state-of-the-art semi-supervised methods.\n\n- [Attentive neural cell instance segmentation](https://www.sciencedirect.com/science/article/pii/S1361841518308442) - Neural cell instance segmentation, which aims at joint detection and segmentation of every neural cell in a microscopic image, is essential to many neuroscience applications. The challenge of this task in- volves cell adhesion, cell distortion, unclear cell contours, low-contrast cell protrusion structures, and background impurities. Consequently, current instance segmentation methods generally fall short of pre- cision. In this paper, we propose an attentive instance segmentation method that accurately predicts the bounding box of each cell as well as its segmentation mask simultaneously. In particular, our method builds on a joint network that combines a single shot multi-box detector (SSD) and a U-net. Furthermore, we employ the attention mechanism in both detection and segmentation modules to focus the model on the useful features. The proposed method is validated on a dataset of neural cell microscopic images. Experimental results demonstrate that our approach can accurately detect and segment neural cell in- stances at a fast speed, comparing favorably with the state-of-the-art methods.\n\n- [LIVECellโA large-scale dataset for label-free live cell segmentation](https://www.nature.com/articles/s41592-021-01249-6) - Light microscopy combined with well-established protocols of two-dimensional cell culture facilitates high-throughput quantitative imaging to study biological phenomena. Accurate segmentation of individual cells in images enables exploration of complex biological questions, but can require sophisticated imaging processing pipelines in cases of low contrast and high object density. Deep learning-based methods are considered state-of-the-art for image segmentation but typically require vast amounts of annotated data, for which there is no suitable resource available in the field of label-free cellular imaging. Here, we present LIVECell, a large, high-quality, manually annotated and expert-validated dataset of phase-contrast images, consisting of over 1.6 million cells from a diverse set of cell morphologies and culture densities. To further demonstrate its use, we train convolutional neural network-based models using LIVECell and evaluate model segmentation accuracy with a proposed a suite of benchmarks. **Thanks to @maxwell110**\n\n- [Nuclei Segmentation via a Deep Panoptic Model with Semantic Feature Fusion](https://www.researchgate.net/publication/334843766_Nuclei_Segmentation_via_a_Deep_Panoptic_Model_with_Semantic_Feature_Fusion) - Automated detection and segmentation of individual nuclei in histopathology images is important for cancer diagnosis and prognosis. Due to the high variability of nuclei appearances and numerous overlapping objects, this task still remains challenging. Deep learning based semantic and instance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent information loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin. **Thanks to @osamurai**\n\n- [Pixel-wise neural cell instance segmentation](https://ieeexplore.ieee.org/abstract/document/8363596) - Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin. **Thanks to @abhishandy**\n\n๐ [U-net with Pytorch](https://github.com/milesial/Pytorch-UNet)\n๐ [Using U-net for a Kaggle competition (AutoEncoder)] (https://github.com/Fpiotro/MOLECULAR-TRANSLATION)\n\n๐ [Model trained by LIVECell for transfer learning] (https://github.com/sartorius-research/LIVECell/tree/main/model) **Thanks to @maxwell110**\n๐ [FIAR's Dectron 2 with Pytorch] (https://github.com/facebookresearch/detectron2) **Thanks to @maxwell110**\n\n**Have a good competition and don't hesitate to comment!**", "votes": null }, { "id": "1548107", "postDate": "10/18/2021 02:09:10", "content": "Thank you for sharing the relevant papers.
\nI think the paper by the organizers of the competition is also quite important, although I think it is a relatively blind spot.
\nEdlund, C., Jackson, T.R., Khalid, N. et al.
\nLIVECell-A large-scale dataset for label-free live cell segmentation.
\nNat Methods 18, 1038 -1045 (2021).
\nAvailable at: https://doi.org/10.1038/s41592-021-01249-6
Their paper is not just a dataset paper, it evaluates the performance of both recent anchor-free and anchor-based models for each cell type under four task conditions, which I think is very informative.
\nWithin their paper, it is shown that the results are better when other cell types not included in the test data are also included in the training. This may be because using even unrelated cell types in the training data allows the model to learn morphology and other factors.
Therefore, in addition to the dataset provided in Kaggle, transfer learning using other cell types in LIVECell, the predecessor of the data, will be important for this competition.
\nUpdate: Oct.19.2021
\n[supplement]
If you want to use the model trained by LIVECell for transfer learning, this page on GitHub may be useful.
\n๐ https://github.com/sartorius-research/LIVECell/tree/main/model
The models are based on PyTorch, but are from FIAR's Dectron 2 library.
\nDectron2: https://github.com/facebookresearch/detectron2
Thank you very much for this reading suggestion ๐
", "rawMarkdown": "Thank you very much for this reading suggestion ๐", "votes": null }, { "id": "1550313", "postDate": "10/19/2021 15:46:45", "content": "Thank you for sharing @datascientistfp
", "rawMarkdown": "Thank you for sharing @datascientistfp", "votes": null }, { "id": "1572836", "postDate": "11/06/2021 02:42:21", "content": "Pixel-wise neural cell instance segmentation: Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin.
", "rawMarkdown": "[Pixel-wise neural cell instance segmentation](https://ieeexplore.ieee.org/abstract/document/8363596): Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin.", "votes": null }, { "id": "1576296", "postDate": "11/09/2021 04:10:53", "content": "Thank you for sharing good summary!
", "rawMarkdown": "Thank you for sharing good summary!", "votes": null }, { "id": "1584475", "postDate": "11/16/2021 15:00:34", "content": "Thank you for creating discussion room for this topic. I also found the paper for not cell but nuclei.
\nBut this paper tackles the problem of overlapping issues. So it may help us. Actually, the proposed model outperforms Mask R-CNN.
\nhttps://www.researchgate.net/publication/334843766_Nuclei_Segmentation_via_a_Deep_Panoptic_Model_with_Semantic_Feature_Fusion
Summary from the paper
\nAutomated detection and segmentation of individ-\nual nuclei in histopathology images is important\nfor cancer diagnosis and prognosis. Due to the\nhigh variability of nuclei appearances and numer-\nous overlapping objects, this task still remains challenging. Deep learning based semantic and in-\nstance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model\n which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent in-formation loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin.\nThank you for sharing!
", "rawMarkdown": "Thank you for sharing!", "votes": null }, { "id": "1599632", "postDate": "11/29/2021 16:17:47", "content": "Thanks for sharing
", "rawMarkdown": "Thanks for sharing", "votes": null }, { "id": "2218048", "postDate": "04/11/2023 11:46:29", "content": "Thank you for sharing!
", "rawMarkdown": "Thank you for sharing!", "votes": null } ], "comments": [ { "id": 1548107, "author_name": "maxwell110", "author_url": "", "post_date": "10/18/2021 02:09:10", "content": "Thank you for sharing the relevant papers.
\nI think the paper by the organizers of the competition is also quite important, although I think it is a relatively blind spot.
\nEdlund, C., Jackson, T.R., Khalid, N. et al.
\nLIVECell-A large-scale dataset for label-free live cell segmentation.
\nNat Methods 18, 1038 -1045 (2021).
\nAvailable at: https://doi.org/10.1038/s41592-021-01249-6
Their paper is not just a dataset paper, it evaluates the performance of both recent anchor-free and anchor-based models for each cell type under four task conditions, which I think is very informative.
\nWithin their paper, it is shown that the results are better when other cell types not included in the test data are also included in the training. This may be because using even unrelated cell types in the training data allows the model to learn morphology and other factors.
Therefore, in addition to the dataset provided in Kaggle, transfer learning using other cell types in LIVECell, the predecessor of the data, will be important for this competition.
\nUpdate: Oct.19.2021
\n[supplement]
If you want to use the model trained by LIVECell for transfer learning, this page on GitHub may be useful.
\n๐ https://github.com/sartorius-research/LIVECell/tree/main/model
The models are based on PyTorch, but are from FIAR's Dectron 2 library.
\nDectron2: https://github.com/facebookresearch/detectron2
Thank you very much for this reading suggestion ๐
", "votes": null, "replies": [] } ] }, { "id": 1550313, "author_name": "raghulsuresh", "author_url": "", "post_date": "10/19/2021 15:46:45", "content": "Thank you for sharing @datascientistfp
", "votes": null, "replies": [] }, { "id": 1572836, "author_name": "abhishandy", "author_url": "", "post_date": "11/06/2021 02:42:21", "content": "Pixel-wise neural cell instance segmentation: Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin.
", "votes": null, "replies": [] }, { "id": 1576296, "author_name": "yoshikuwano", "author_url": "", "post_date": "11/09/2021 04:10:53", "content": "Thank you for sharing good summary!
", "votes": null, "replies": [] }, { "id": 1584475, "author_name": "osamurai", "author_url": "", "post_date": "11/16/2021 15:00:34", "content": "Thank you for creating discussion room for this topic. I also found the paper for not cell but nuclei.
\nBut this paper tackles the problem of overlapping issues. So it may help us. Actually, the proposed model outperforms Mask R-CNN.
\nhttps://www.researchgate.net/publication/334843766_Nuclei_Segmentation_via_a_Deep_Panoptic_Model_with_Semantic_Feature_Fusion
Summary from the paper
\nAutomated detection and segmentation of individ-\nual nuclei in histopathology images is important\nfor cancer diagnosis and prognosis. Due to the\nhigh variability of nuclei appearances and numer-\nous overlapping objects, this task still remains challenging. Deep learning based semantic and in-\nstance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model\n which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent in-formation loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin.\nThank you for sharing!
", "votes": null, "replies": [] }, { "id": 1599632, "author_name": "ahmedtarek26", "author_url": "", "post_date": "11/29/2021 16:17:47", "content": "Thanks for sharing
", "votes": null, "replies": [] }, { "id": 2218048, "author_name": "abdelouahabzaari", "author_url": "", "post_date": "04/11/2023 11:46:29", "content": "Thank you for sharing!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1545410": "Hello everyone!\n\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I downloaded the articles that seemed relevant after reading their summaries.\n\n**Papers:**\n\n- [Instance Segmentation of Neural Cells](https://www.researchgate.net/publication/329933921_Instance_Segmentation_of_Neural_Cells) - Instance segmentation of neural cells plays an important role in brain study. However, this task is challenging due to the special shapes and behaviors of neural cells. Existing methods are not precise enough to capture their tiny structures, e.g., filopodia and lamellipodia, which are critical to the understanding of cell interaction and behavior. To this end, we propose a novel deep multi-task learning model to jointly detect and segment neural cells instance-wise. Our method is built upon SSD, with ResNet101 as the backbone to achieve both high detection accuracy and fast speed. Furthermore, unlike existing works which tend to produce wavy and inaccurate boundaries, we embed a deconvolution module into SSD to better capture details. Experiments on a dataset of neural cell microscopic images show that our method is able to achieve better performance in terms of accuracy and efficiency, comparing favorably with current state-of-the-art methods.\n\n- [Mask-RCNN for Cell Instance Segmentation](https://www.researchgate.net/publication/340998089_Rank_6_Mask-RCNN_for_Cell_Instance_Segmentation) - We proposed an automatic nucleus segmentation algorithm of H&E stained tissue microscopy images. Mask-RCNN is a recently proposed state-of-the-art algorithm for object detection and object instance segmentation of natural images. In this paper, we demonstrate that Mask-RCNN can be used to perform highly effective and efficient automatic segmentation of H&E microscopy images for cell nuclei. We propose a novel MASK Non-maximum suppression (NMS) module which can automatically ensemble classifiers results and increase the robustness of model.\n\n- [IRNet: Instance Relation Network for\nOverlapping Cervical Cell Segmentation](https://arxiv.org/abs/1908.06623) - Cell instance segmentation in Pap smear image remains challenging due to the wide existence of occlusion among translucent cytoplasm in cell clumps. Conventional methods heavily rely on accurate nuclei detection results and are easily disturbed by miscellaneous objects. In this paper, we propose a novel Instance Relation Network (IRNet) for robust overlapping cell segmentation by exploring instance relation interaction. Specifically, we propose the Instance Relation Module to construct the cell association matrix for transferring information among individual cell-instance features. With the collaboration of different instances, the augmented features gain benefits from contextual information and improve semantic consistency. Meanwhile, we proposed a sparsity constrained Duplicate Removal Module to eliminate the misalignment between classification and localization accuracy for candidates selection. The largest cervical Pap smear (CPS) dataset with more than 8000 cell annotations in Pap smear image was constructed for comprehensive evaluation. Our method outperforms other methods by a large margin, demonstrating the effectiveness of exploring instance relation. \n\n- [Weakly Supervised Cell Instance Segmentation by Propagating from Detection Response](https://www.researchgate.net/publication/336392797_Weakly_Supervised_Cell_Instance_Segmentation_by_Propagating_from_Detection_Response) - Cell shape analysis is important in biomedical research. Deep learning methods may perform to segment individual cells if they use sufficient training data that the boundary of each cell is annotated. However, it is very time-consuming for preparing such detailed annotation for many cell culture conditions. In this paper, we propose a weakly supervised method that can segment individual cell regions who touch each other with unclear boundaries in dense conditions without the training data for cell regions. We demonstrated the efficacy of our method using several data-set including multiple cell types captured by several types of microscopy. Our method achieved the highest accuracy compared with several conventional methods. In addition, we demonstrated that our method can perform without any annotation by using fluorescence images that cell nuclear were stained as training data.\n\n- [Deep Semi-supervised Knowledge Distillation for\nOverlapping Cervical Cell Instance Segmentation](https://arxiv.org/abs/2007.10787) - Deep learning methods show promising results for overlapping cervical cell instance segmentation. However, in order to train a model with good generalization ability, voluminous pixel-level annotations are demanded which is quite expensive and time-consuming for acquisition. In this paper, we propose to leverage both labeled and unlabeled data for instance segmentation with improved accuracy by knowledge distillation. We propose a novel Mask-guided Mean Teacher framework with Perturbation-sensitive Sample Mining (MMT-PSM), which consists of a teacher and a student network during training. Two networks are encouraged to be consistent both in feature and semantic level under small perturbations. The teacherโs self-ensemble predictions from K-time augmented samples are used to construct the reliable pseudolabels for optimizing the student. We design a novel strategy to estimate the sensitivity to perturbations for each proposal and select informative samples from massive cases to facilitate fast and effective semantic distillation. In addition, to eliminate the unavoidable noise from the background region, we propose to use the predicted segmentation mask as guidance to enforce the feature distillation in the foreground region. Experiments show that the proposed method improves the performance significantly compared with the supervised method learned from labeled data only, and outperforms state-of-the-art semi-supervised methods.\n\n- [Attentive neural cell instance segmentation](https://www.sciencedirect.com/science/article/pii/S1361841518308442) - Neural cell instance segmentation, which aims at joint detection and segmentation of every neural cell in a microscopic image, is essential to many neuroscience applications. The challenge of this task in- volves cell adhesion, cell distortion, unclear cell contours, low-contrast cell protrusion structures, and background impurities. Consequently, current instance segmentation methods generally fall short of pre- cision. In this paper, we propose an attentive instance segmentation method that accurately predicts the bounding box of each cell as well as its segmentation mask simultaneously. In particular, our method builds on a joint network that combines a single shot multi-box detector (SSD) and a U-net. Furthermore, we employ the attention mechanism in both detection and segmentation modules to focus the model on the useful features. The proposed method is validated on a dataset of neural cell microscopic images. Experimental results demonstrate that our approach can accurately detect and segment neural cell in- stances at a fast speed, comparing favorably with the state-of-the-art methods.\n\n- [LIVECellโA large-scale dataset for label-free live cell segmentation](https://www.nature.com/articles/s41592-021-01249-6) - Light microscopy combined with well-established protocols of two-dimensional cell culture facilitates high-throughput quantitative imaging to study biological phenomena. Accurate segmentation of individual cells in images enables exploration of complex biological questions, but can require sophisticated imaging processing pipelines in cases of low contrast and high object density. Deep learning-based methods are considered state-of-the-art for image segmentation but typically require vast amounts of annotated data, for which there is no suitable resource available in the field of label-free cellular imaging. Here, we present LIVECell, a large, high-quality, manually annotated and expert-validated dataset of phase-contrast images, consisting of over 1.6 million cells from a diverse set of cell morphologies and culture densities. To further demonstrate its use, we train convolutional neural network-based models using LIVECell and evaluate model segmentation accuracy with a proposed a suite of benchmarks. **Thanks to @maxwell110**\n\n- [Nuclei Segmentation via a Deep Panoptic Model with Semantic Feature Fusion](https://www.researchgate.net/publication/334843766_Nuclei_Segmentation_via_a_Deep_Panoptic_Model_with_Semantic_Feature_Fusion) - Automated detection and segmentation of individual nuclei in histopathology images is important for cancer diagnosis and prognosis. Due to the high variability of nuclei appearances and numerous overlapping objects, this task still remains challenging. Deep learning based semantic and instance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent information loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin. **Thanks to @osamurai**\n\n- [Pixel-wise neural cell instance segmentation](https://ieeexplore.ieee.org/abstract/document/8363596) - Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin. **Thanks to @abhishandy**\n\n๐ [U-net with Pytorch](https://github.com/milesial/Pytorch-UNet)\n๐ [Using U-net for a Kaggle competition (AutoEncoder)] (https://github.com/Fpiotro/MOLECULAR-TRANSLATION)\n\n๐ [Model trained by LIVECell for transfer learning] (https://github.com/sartorius-research/LIVECell/tree/main/model) **Thanks to @maxwell110**\n๐ [FIAR's Dectron 2 with Pytorch] (https://github.com/facebookresearch/detectron2) **Thanks to @maxwell110**\n\n**Have a good competition and don't hesitate to comment!**", "1548107": "Thank you for sharing the relevant papers.\n\nI think the paper by the organizers of the competition is also quite important, although I think it is a relatively blind spot.\n\nEdlund, C., Jackson, T.R., Khalid, N. et al. \nLIVECell-A large-scale dataset for label-free live cell segmentation. \nNat Methods 18, 1038 -1045 (2021). \nAvailable at: https://doi.org/10.1038/s41592-021-01249-6\n\nTheir paper is not just a dataset paper, it evaluates the performance of both recent anchor-free and anchor-based models for each cell type under four task conditions, which I think is very informative.\nWithin their paper, it is shown that `the results are better when other cell types not included in the test data are also included in the training`. This may be because using even unrelated cell types in the training data allows the model to learn morphology and other factors.\n\nTherefore, in addition to the dataset provided in Kaggle, transfer learning using other cell types in LIVECell, the predecessor of the data, will be important for this competition.\n\n---\nUpdate: Oct.19.2021\n[supplement]\n\nIf you want to use the model trained by LIVECell for transfer learning, this page on GitHub may be useful.\n๐ https://github.com/sartorius-research/LIVECell/tree/main/model\n\nThe models are based on PyTorch, but are from FIAR's Dectron 2 library.\nDectron2: https://github.com/facebookresearch/detectron2", "1549004": "Thank you very much for this reading suggestion ๐", "1550313": "Thank you for sharing @datascientistfp", "1572836": "[Pixel-wise neural cell instance segmentation](https://ieeexplore.ieee.org/abstract/document/8363596): Accurate cell instance segmentation plays an important role in the study of neural cell interactions, which are critical for understanding the development of brain. These interactions are performed through the filopodia and lamellipodia of neural cells, which are extremely tiny structures and as a result render most existing instance segmentation methods powerless to precisely capture them. To solve this issue, in this paper we present a novel hierarchical neural network comprising object detection and segmentation modules. Compared to previous work, our model is able to efficiently share and make full use of the information at different levels between the two modules. Our method is simple yet powerful, and experimental results show that it captures the contours of neural cells, especially the filopodia and lamellipodia, with high accuracy, and outperforms recent state of the art by a large margin.", "1576296": "Thank you for sharing good summary!", "1584475": "Thank you for creating discussion room for this topic. I also found the paper for not cell but nuclei.\nBut this paper tackles the problem of overlapping issues. So it may help us. Actually, the proposed model outperforms Mask R-CNN.\nhttps://www.researchgate.net/publication/334843766_Nuclei_Segmentation_via_a_Deep_Panoptic_Model_with_Semantic_Feature_Fusion\n\nSummary from the paper\n```\nAutomated detection and segmentation of individ-\nual nuclei in histopathology images is important\nfor cancer diagnosis and prognosis. Due to the\nhigh variability of nuclei appearances and numer-\nous overlapping objects, this task still remains challenging. Deep learning based semantic and in-\nstance segmentation models have been proposed to address the challenges, but these methods tend to concentrate on either the global or local features and hence still suffer from information loss. In this work, we propose a panoptic segmentation model\n which incorporates an auxiliary semantic segmentation branch with the instance branch to integrate global and local features. Furthermore, we design a feature map fusion mechanism in the instance branch and a new mask generator to prevent in-formation loss. Experimental results on three different histopathology datasets demonstrate that our method outperforms the state-of-the-art nuclei segmentation methods and popular semantic and instance segmentation models by a large margin.\n```", "1587259": "Thank you for sharing!", "1599632": "Thanks for sharing", "2218048": "Thank you for sharing!" }, "source": "meta" }