The application of Mask R-CNNs (Regional Convolutional Neural Networks) is being explored for solving instance segmentation problems. The primary focus of this project is on creating an edge model that will operate fully on mobile or other edge devices, aimed at educational and learning purposes. More importantly, the goal is to develop an AI model that will streamline and enhance the human application and understanding of microscopy slides, enabling more curated diagnoses. In developing countries, especially in remote areas, the complexity of these activities often makes management challenging, particularly when technical expertise is limited.
- Pytorch
- Lightning: A third party library built on top of Pytorch which aims to hyper-resonate the speed of Pytorch models and tasks
- torchvision: A pytorch lib dedicated to computer vision tasks
- Albumentations: A third party lib for image and bounding box augmentations
- Libraries like orjson, sklearn, matplotlib, seaborn
- cv2: a computer vision library
The project is currently under development.