This project demonstrates how to use U-Net for automating nucleus detection as part of the 2018 Data Science Bowl. The goal is to facilitate faster research into various diseases by improving nucleus detection accuracy in cell images.
The Jupyter Notebook in this repository provides a detailed implementation of U-Net for semantic segmentation of cell nuclei. The notebook walks through the process of combining mask images, training the U-Net model, and performing nucleus detection on test images.
Semantic_Segmentation_using_UNET.ipynb: The Jupyter Notebook containing the implementation of U-Net for nucleus detection.
- Kaggle platform - used for free GPU
- U-net for semantic segmentation
- Dataset - 2018 Data Science Bowl(kaggle)
Make sure you have the following installed:
- Python 3.x
- TensorFlow
- Numpy
- Skimage (for image processing)
- Matplotlib (plotting the images and graphs)