![\"1\"](\"http://brians.network/images/plots.png\")
What sort of cool looking or interesting looking pictures are you making out there? I visualized some of the augmentation and manipulations I've been trying recently and came up with these cool looking images:
\n\n
What sort of cool looking or interesting looking pictures are you making out there? I visualized some of the augmentation and manipulations I've been trying recently and came up with these cool looking images:
\n\n
Cool. How did you do that?
", "rawMarkdown": "Cool. How did you do that?", "votes": null }, { "id": "437948", "postDate": "12/12/2018 20:31:34", "content": "I am using the idea from @alexijie to separate the data into cell crops. The problem is that each crop may not have all targets. These images are combined from all the individual crops for each sample.
\n\nI've put together a generator with cv2 and I am trying things like histogram equalization, averaging the samples, doing a running average of the image. This image came from applying histogram equalization on the RGBY channels individually and then averaging all of the cells together.
", "rawMarkdown": "I am using the idea from @alexijie to separate the data into cell crops. The problem is that each crop may not have all targets. These images are combined from all the individual crops for each sample. \n\nI've put together a generator with cv2 and I am trying things like histogram equalization, averaging the samples, doing a running average of the image. This image came from applying histogram equalization on the RGBY channels individually and then averaging all of the cells together.", "votes": null }, { "id": "438353", "postDate": "12/13/2018 14:30:38", "content": "@Brian what technique are you using the create the cell crops?
", "rawMarkdown": "Brian what technique are you using the create the cell crops?", "votes": null }, { "id": "438426", "postDate": "12/13/2018 16:36:25", "content": "I am using the technique that @alexijie demonstrates. I modified the code to change the segmentation parameters, run with multiprocessing, and to combine the images in different ways:\n https://github.com/alexxijielu/paired_cell_inpainting
", "rawMarkdown": "I am using the technique that @alexijie demonstrates. I modified the code to change the segmentation parameters, run with multiprocessing, and to combine the images in different ways:\n https://github.com/alexxijielu/paired_cell_inpainting", "votes": null }, { "id": "439027", "postDate": "12/14/2018 16:06:22", "content": "I wonder if living near a dispensary in Colorado would provide additional visual augmentation.
", "rawMarkdown": "I wonder if living near a dispensary in Colorado would provide additional visual augmentation.", "votes": null } ], "comments": [ { "id": 437672, "author_name": "garybios", "author_url": "", "post_date": "12/12/2018 09:50:24", "content": "Cool. How did you do that?
", "votes": null, "replies": [ { "id": 437948, "author_name": "ldm314", "author_url": "", "post_date": "12/12/2018 20:31:34", "content": "I am using the idea from @alexijie to separate the data into cell crops. The problem is that each crop may not have all targets. These images are combined from all the individual crops for each sample.
\n\nI've put together a generator with cv2 and I am trying things like histogram equalization, averaging the samples, doing a running average of the image. This image came from applying histogram equalization on the RGBY channels individually and then averaging all of the cells together.
", "votes": null, "replies": [] } ] }, { "id": 438353, "author_name": "hortonhearsafoo", "author_url": "", "post_date": "12/13/2018 14:30:38", "content": "@Brian what technique are you using the create the cell crops?
", "votes": null, "replies": [ { "id": 438426, "author_name": "ldm314", "author_url": "", "post_date": "12/13/2018 16:36:25", "content": "I am using the technique that @alexijie demonstrates. I modified the code to change the segmentation parameters, run with multiprocessing, and to combine the images in different ways:\n https://github.com/alexxijielu/paired_cell_inpainting
", "votes": null, "replies": [] } ] }, { "id": 439027, "author_name": "sasrdw", "author_url": "", "post_date": "12/14/2018 16:06:22", "content": "I wonder if living near a dispensary in Colorado would provide additional visual augmentation.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "437530": "What sort of cool looking or interesting looking pictures are you making out there? I visualized some of the augmentation and manipulations I've been trying recently and came up with these cool looking images:\n\n ", "437672": "Cool. How did you do that?", "437948": "I am using the idea from @alexijie to separate the data into cell crops. The problem is that each crop may not have all targets. These images are combined from all the individual crops for each sample. \n\nI've put together a generator with cv2 and I am trying things like histogram equalization, averaging the samples, doing a running average of the image. This image came from applying histogram equalization on the RGBY channels individually and then averaging all of the cells together.", "438353": "Brian what technique are you using the create the cell crops?", "438426": "I am using the technique that @alexijie demonstrates. I modified the code to change the segmentation parameters, run with multiprocessing, and to combine the images in different ways:\n https://github.com/alexxijielu/paired_cell_inpainting", "439027": "I wonder if living near a dispensary in Colorado would provide additional visual augmentation." }, "source": "meta" }