# UrineCellSegmentation **Repository Path**: suppermanljr/UrineCellSegmentation ## Basic Information - **Project Name**: UrineCellSegmentation - **Description**: No description available - **Primary Language**: Unknown - **License**: Not specified - **Default Branch**: main - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2026-11-18 - **Last Updated**: 2026-11-18 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Detectron for Urine Cell Segmentation # Scripts * `preprocess.py` * splits up the segmentation and mask portion * `convert_coco.py` * contains the utility function for creating a JSON in the COCO data format * Resources for understanding COCO * https://detectron2.readthedocs.io/en/latest/tutorials/datasets.html * https://www.immersivelimit.com/tutorials/create-coco-annotations-from-scratch/#coco-dataset-format * Methods 1. Identify all connected components of the same color * findComponents(class): returns a mask for each connected region of given color (the class) * https://scipy-lectures.org/packages/scikit-image/auto_examples/plot_labels.html 2. Convert that to run length encodings * `pycocotools.mask.encode(np.asarray(mask, order="F"))` * `training.py` * trains the model * `gen_output.py` * creates some images with the segmentation visualization overlayed # Other Resources [Data Source](https://github.com/jlevy44/PreliminaryGenerativeHistoPath/) [Detectron Tutorial](https://colab.research.google.com/drive/16jcaJoc6bCFAQ96jDe2HwtXj7BMD_-m5#scrollTo=U5LhISJqWXgM)