Status: Active (updates are to be expected)

Bio-inspired Model Image Segmentation Benchmark Documentation

Prerequisites

Install Git Repository

git clone https://github.com/rainarit/segmentation_benchmark.git
cd segmentation_benchmark

Create Conda Environment

The code in the tutorial has been written using Python 3; many of the dependencies may not be available for Python 2.7.

We strongly recommend using the Anaconda Python distribution. You can install either the full anaconda distribution (very extensive, but large) or miniconda (much smaller, only essential packages).

Create a new conda environment with the use of the package-list.txt

conda create --name <env> --file package-list.txt

Run Training Script(s):

Download COCO Dataset:

./coco_download.sh

Run Training Script:

./train_fcn_resnet50_coco.sh

License

This project is licensed under the MIT License - see the LICENSE.md file for details