\"Introduction to U-Net and Res-Net for Image Segmentation\" By Aditi Mittal
\n\"Computer sees the images as matrices which need to be processed to get a meaning out of it.\"
\n\"Image segmentation is the method to partition the image into various segments with each segment having a different entity. Convolutional Neural Networks are successful for simpler images but haven’t given good results for complex images. This is where other algorithms like U-Net and Res-Net come into play.\"
\"CNNs are similar to a neural network with various neutrons with learnable weights and biases. Each neuron is given a number of inputs, weighted sum is performed, activation function is applied and output is given. The network has a loss function which is used to minimize the error in weights.\"
\n\"A machine sees an image as a matrix of pixels with image resolution as h x w x d where h is the height, w is the width and d is the dimension. d depends on the color scale such as 3 for RGB scale and 1 for grayscale.\"
\n\"In CNN, the image is converted into a vector which is largely used in classification problems. But in U-Net, an image is converted into a vector and then the same mapping is used to convert it again to an image. This reduces the distortion by preserving the original structure of the image.\"
\n\"CNN is largely used when the whole image is needed to be classified as a class label. But many tasks requires to classify each pixel of the image. This is solved by the U-net and Res-Net.\"
\n\"U-Net consists of Convolution Operation, Max Pooling, ReLU Activation, Concatenation and Up Sampling Layers and three sections: contraction, bottleneck, and expansion section. U-net uses a loss function for each pixel of the image. This helps in easy identification of individual cells within the segmentation map. Softmax is applied to each pixel followed by a loss function. This converts the segmentation problem into a classification problem where we need to classify each pixel to one of the classes.\"
\n\"In traditional neural networks, more layers mean a better network but because of the vanishing gradient problem, weights of the first layer won’t be updated correctly through the back-propagation. As the error gradient is back-propagated to earlier layers, repeated multiplication makes the gradient small. Thus, with more layers in the networks, its performance gets saturated and starts decreasing rapidly.\"
\n\" Res-Net solves this problem by using the identity matrix. When the back-propagation is done through identity function, the gradient will be multiplied only by 1. This preserves the input and avoids any loss in the information.\"
\n", "messages": [ { "id": "1731877", "postDate": "03/22/2022 19:45:38", "content": "\"Introduction to U-Net and Res-Net for Image Segmentation\" By Aditi Mittal
\n\"Computer sees the images as matrices which need to be processed to get a meaning out of it.\"
\n\"Image segmentation is the method to partition the image into various segments with each segment having a different entity. Convolutional Neural Networks are successful for simpler images but haven’t given good results for complex images. This is where other algorithms like U-Net and Res-Net come into play.\"
\"CNNs are similar to a neural network with various neutrons with learnable weights and biases. Each neuron is given a number of inputs, weighted sum is performed, activation function is applied and output is given. The network has a loss function which is used to minimize the error in weights.\"
\n\"A machine sees an image as a matrix of pixels with image resolution as h x w x d where h is the height, w is the width and d is the dimension. d depends on the color scale such as 3 for RGB scale and 1 for grayscale.\"
\n\"In CNN, the image is converted into a vector which is largely used in classification problems. But in U-Net, an image is converted into a vector and then the same mapping is used to convert it again to an image. This reduces the distortion by preserving the original structure of the image.\"
\n\"CNN is largely used when the whole image is needed to be classified as a class label. But many tasks requires to classify each pixel of the image. This is solved by the U-net and Res-Net.\"
\n\"U-Net consists of Convolution Operation, Max Pooling, ReLU Activation, Concatenation and Up Sampling Layers and three sections: contraction, bottleneck, and expansion section. U-net uses a loss function for each pixel of the image. This helps in easy identification of individual cells within the segmentation map. Softmax is applied to each pixel followed by a loss function. This converts the segmentation problem into a classification problem where we need to classify each pixel to one of the classes.\"
\n\"In traditional neural networks, more layers mean a better network but because of the vanishing gradient problem, weights of the first layer won’t be updated correctly through the back-propagation. As the error gradient is back-propagated to earlier layers, repeated multiplication makes the gradient small. Thus, with more layers in the networks, its performance gets saturated and starts decreasing rapidly.\"
\n\" Res-Net solves this problem by using the identity matrix. When the back-propagation is done through identity function, the gradient will be multiplied only by 1. This preserves the input and avoids any loss in the information.\"
\n", "rawMarkdown": "\"Introduction to U-Net and Res-Net for Image Segmentation\" By Aditi Mittal\n\n\"Computer sees the images as matrices which need to be processed to get a meaning out of it.\"\n\"Image segmentation is the method to partition the image into various segments with each segment having a different entity. Convolutional Neural Networks are successful for simpler images but haven’t given good results for complex images. This is where other algorithms like U-Net and Res-Net come into play.\"\n\n#CNN (Convolutional Neural Networks)\n\n\"CNNs are similar to a neural network with various neutrons with learnable weights and biases. Each neuron is given a number of inputs, weighted sum is performed, activation function is applied and output is given. The network has a loss function which is used to minimize the error in weights.\"\n\n\"A machine sees an image as a matrix of pixels with image resolution as h x w x d where h is the height, w is the width and d is the dimension. d depends on the color scale such as 3 for RGB scale and 1 for grayscale.\"\n\n\"In CNN, the image is converted into a vector which is largely used in classification problems. But in U-Net, an image is converted into a vector and then the same mapping is used to convert it again to an image. This reduces the distortion by preserving the original structure of the image.\"\n\n\"CNN is largely used when the whole image is needed to be classified as a class label. But many tasks requires to classify each pixel of the image. This is solved by the U-net and Res-Net.\"\n\n#U-Net\n\n\"U-Net consists of Convolution Operation, Max Pooling, ReLU Activation, Concatenation and Up Sampling Layers and three sections: contraction, bottleneck, and expansion section. U-net uses a loss function for each pixel of the image. This helps in easy identification of individual cells within the segmentation map. Softmax is applied to each pixel followed by a loss function. This converts the segmentation problem into a classification problem where we need to classify each pixel to one of the classes.\"\n\n#Residual Networks (Res-Net)\n\n\"In traditional neural networks, more layers mean a better network but because of the vanishing gradient problem, weights of the first layer won’t be updated correctly through the back-propagation. As the error gradient is back-propagated to earlier layers, repeated multiplication makes the gradient small. Thus, with more layers in the networks, its performance gets saturated and starts decreasing rapidly.\"\n\n\" Res-Net solves this problem by using the identity matrix. When the back-propagation is done through identity function, the gradient will be multiplied only by 1. This preserves the input and avoids any loss in the information.\"\n\nhttps://aditi-mittal.medium.com/introduction-to-u-net-and-res-net-for-image-segmentation-9afcb432ee2f", "votes": null }, { "id": "1731942", "postDate": "03/22/2022 22:06:39", "content": "Nice work！
", "rawMarkdown": "Nice work！", "votes": null }, { "id": "1731951", "postDate": "03/22/2022 22:13:13", "content": "Thank you Alan. I'm trying to understand what this competition is about:) And learn a little bit more.
", "rawMarkdown": "Thank you Alan. I'm trying to understand what this competition is about:) And learn a little bit more.", "votes": null }, { "id": "1732017", "postDate": "03/22/2022 23:54:43", "content": "Really helpful!
\nThank you!
Thank you Thales. As I wrote above, there are so many things to learn with those subjects during this competition.
", "rawMarkdown": "Thank you Thales. As I wrote above, there are so many things to learn with those subjects during this competition.", "votes": null }, { "id": "1743883", "postDate": "04/03/2022 11:45:58", "content": "Super cool!
\nThank you
Thank you Mohammed. Your words mean a lot to me.
", "rawMarkdown": "Thank you Mohammed. Your words mean a lot to me.", "votes": null }, { "id": "1766244", "postDate": "04/24/2022 10:36:45", "content": "Thank you!!
", "rawMarkdown": "Thank you!!", "votes": null }, { "id": "1766387", "postDate": "04/24/2022 14:08:20", "content": "Congrats for your 29th place in this Hotel ID Human traffic competition machikomomo.
", "rawMarkdown": "Congrats for your 29th place in this Hotel ID Human traffic competition machikomomo.", "votes": null }, { "id": "2681562", "postDate": "03/04/2024 17:51:10", "content": "Thanks, I cleared my doubts from this topic.
", "rawMarkdown": "Thanks, I cleared my doubts from this topic.", "votes": null } ], "comments": [ { "id": 1731942, "author_name": "alanhabrony", "author_url": "", "post_date": "03/22/2022 22:06:39", "content": "Nice work！
", "votes": null, "replies": [ { "id": 1731951, "author_name": "mpwolke", "author_url": "", "post_date": "03/22/2022 22:13:13", "content": "Thank you Alan. I'm trying to understand what this competition is about:) And learn a little bit more.
", "votes": null, "replies": [] } ] }, { "id": 1732017, "author_name": "thalesgomes", "author_url": "", "post_date": "03/22/2022 23:54:43", "content": "Really helpful!
\nThank you!
Thank you Thales. As I wrote above, there are so many things to learn with those subjects during this competition.
", "votes": null, "replies": [] } ] }, { "id": 1743883, "author_name": "mohammeda1i", "author_url": "", "post_date": "04/03/2022 11:45:58", "content": "Super cool!
\nThank you
Thank you Mohammed. Your words mean a lot to me.
", "votes": null, "replies": [] } ] }, { "id": 1766244, "author_name": "machikomomo", "author_url": "", "post_date": "04/24/2022 10:36:45", "content": "Thank you!!
", "votes": null, "replies": [ { "id": 1766387, "author_name": "mpwolke", "author_url": "", "post_date": "04/24/2022 14:08:20", "content": "Congrats for your 29th place in this Hotel ID Human traffic competition machikomomo.
", "votes": null, "replies": [] } ] }, { "id": 2681562, "author_name": "amarnathreddys", "author_url": "", "post_date": "03/04/2024 17:51:10", "content": "Thanks, I cleared my doubts from this topic.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1731877": "\"Introduction to U-Net and Res-Net for Image Segmentation\" By Aditi Mittal\n\n\"Computer sees the images as matrices which need to be processed to get a meaning out of it.\"\n\"Image segmentation is the method to partition the image into various segments with each segment having a different entity. Convolutional Neural Networks are successful for simpler images but haven’t given good results for complex images. This is where other algorithms like U-Net and Res-Net come into play.\"\n\n#CNN (Convolutional Neural Networks)\n\n\"CNNs are similar to a neural network with various neutrons with learnable weights and biases. Each neuron is given a number of inputs, weighted sum is performed, activation function is applied and output is given. The network has a loss function which is used to minimize the error in weights.\"\n\n\"A machine sees an image as a matrix of pixels with image resolution as h x w x d where h is the height, w is the width and d is the dimension. d depends on the color scale such as 3 for RGB scale and 1 for grayscale.\"\n\n\"In CNN, the image is converted into a vector which is largely used in classification problems. But in U-Net, an image is converted into a vector and then the same mapping is used to convert it again to an image. This reduces the distortion by preserving the original structure of the image.\"\n\n\"CNN is largely used when the whole image is needed to be classified as a class label. But many tasks requires to classify each pixel of the image. This is solved by the U-net and Res-Net.\"\n\n#U-Net\n\n\"U-Net consists of Convolution Operation, Max Pooling, ReLU Activation, Concatenation and Up Sampling Layers and three sections: contraction, bottleneck, and expansion section. U-net uses a loss function for each pixel of the image. This helps in easy identification of individual cells within the segmentation map. Softmax is applied to each pixel followed by a loss function. This converts the segmentation problem into a classification problem where we need to classify each pixel to one of the classes.\"\n\n#Residual Networks (Res-Net)\n\n\"In traditional neural networks, more layers mean a better network but because of the vanishing gradient problem, weights of the first layer won’t be updated correctly through the back-propagation. As the error gradient is back-propagated to earlier layers, repeated multiplication makes the gradient small. Thus, with more layers in the networks, its performance gets saturated and starts decreasing rapidly.\"\n\n\" Res-Net solves this problem by using the identity matrix. When the back-propagation is done through identity function, the gradient will be multiplied only by 1. This preserves the input and avoids any loss in the information.\"\n\nhttps://aditi-mittal.medium.com/introduction-to-u-net-and-res-net-for-image-segmentation-9afcb432ee2f", "1731942": "Nice work！", "1731951": "Thank you Alan. I'm trying to understand what this competition is about:) And learn a little bit more.", "1732017": "Really helpful!\nThank you!", "1732446": "Thank you Thales. As I wrote above, there are so many things to learn with those subjects during this competition.", "1743883": "Super cool!\nThank you", "1743990": "Thank you Mohammed. Your words mean a lot to me.", "1766244": "Thank you!!", "1766387": "Congrats for your 29th place in this Hotel ID Human traffic competition machikomomo.", "2681562": "Thanks, I cleared my doubts from this topic." }, "source": "meta" }