Fine-tuning a pre-trained object detection model for the Standford Drone Dataset.

Code in sdd-utils comes from here.

• Requires ffmpeg to be installed.

pip install -r requirements.txt

and for to create images from video frames and create Pascal VOC dataset.

sdd-utils/annotate.py

cd pytorch

pip install -r requirements.txt

python src/utils_.py

python src/main.py