School in AI Project Work

This repository contains the code to train and evaluate a pedestrian detector for the "School in Ai 2° edition"@@UNIMORE

Installation

N.B.: Installation only avaiable in win64 environments

Create and activate an environment with all required packages:

conda create --name ped_detector --file deps/wins/conda_environment.txt
# or conda env create -f deps/win/conda_environment.yml
conda activate cvcspw
pip install -r deps/win/pip_requirements.txt

Dataset download and preparation:

Solution 1 - From Google Drice

Download the storage folder directly from Google Drive [here](link google drive) and place it in the root dir of the project After runnning this step, your storage directory should look like this:

storage
    ├── MOTChallenge
        ├── MOT17
        ├── motcha_coco_annotations
    ├── MOTSynth
        ├── annotations
        ├── comb_annotations
        ├── frames
    ├── motsynth_output

Solution 2 - From scratch

Prepare MOTSynth dataset

Download MOTSynth_1.

wget -P ./storage/MOTSynth https://motchallenge.net/data/MOTSynth_1.zip
unzip ./storage/MOTSynth/MOTSynth_1.zip
rm ./storage/MOTSynth/MOTSynth_1.zip

Delete video from 123 to 256
Extract frames from the videos

python tools/anns/to_frames.py --motsynth-root ./storage/MOTSynth

# now you can delete other videos
rm -r ./storage/MOTSynth/MOTSynth_1

Download and extract annotations

wget -P ./storage/MOTSynth https://motchallenge.net/data/MOTSynth_coco_annotations.zip
unzip ./storage/MOTSynth/MOTSynth_coco_annotations.zip
rm ./storage/MOTSynth/MOTSynth_coco_annotations.zip

Prepare combined annotations for MOTSynth from the original coco annotations

python tools/anns/combine_anns.py --motsynth-path ./storage/MOTSynth

Prepare MOT17 dataset

Colab Usage

You can also use Google Colab if you need remote resources like GPUs. In the notebook folder you can find some useful .ipynb files and remember to load the storage folder in your GDrive before usage.

Object Detection

An adaption of torchvision's detection reference code is done to train Faster R-CNN on a portion of the MOTSynth dataset. To train the model you can run:

./scripts/train_detector