MLNS-Final-Project

Report

In this project, we explore several approaches to optimise target re-identification (re-ID) as a re-ranking problem. Our work consisted in trying different methods to re-rank the re-ID results.

Datasets

We will be testing our work on Market-1501 and VeriWild

Requirements

• dgl
• Pytorch
• scikit-learn

GNN based reranking

The code has been included in /extension. To compile it:

cd extension
sh make.sh

To run reranking evaluation:

1. Place dataset files under 'dataset/' folder: The dataset structure should be like:

datasets/
    Market/
        camids.pkl
        feat.pkl
        ids.pkl

2. python run.py Ps: Reranking runs only on GPU

Members

Chady Raach Mehdi Zemni Marah Gamdou