1. Developed a vehicle detection model to analyze traffic flow at key junctions in Jodhpur, using CCTV footage from IIT Jodhpur’s Department of Civil Engineering.

  2. Trained state-of-the-art models like YOLOv3 on custom datasets tailored to Indian vehicles, including auto rickshaws and tempos.

  3. This model was then used as the first stage of a larger vehicular trajectory prediction model.

Model Input

Model Input This is the snapshot of the raw CCTV footage on which the classification was to be done.

Model Output

Model Output This is the output of the classification model which was trained on a custom dataset.