This project has been implemented for the purpose of learning and applying the main concepts of object detection (multi-class classification + localization using bounding boxes)
!! Important - Numpy 1.17 must be installed (not later version) !!
Clone the Repo:
git clone https://github.com/kosletr/Food-Object-Detection-Pytorch-FasterRCNN/
Install Requirements by opening a terminal and running:
pip install -r requirements.txt
Download the UECFOOD100 dataset from http://foodcam.mobi/dataset100.html and extract it.
Open terminal and run the script to start the training process:
python mainObjDet.py
After 10 epochs of training, some of the results produced by the model are shown below:
Credits to pytorch.org tutorial found in here:
http://pytorch.org/tutorials/intermediate/torchvision_tutorial.html
and the creators of the dataset at: