predict-vehicle-speed-comma-ai-challenge

Following is the implementation of DeepVO architecture for predicting vehicle speed from a Video of Dash Camera.

Source description

├─── data-pre-processing/ : Source responsible for data-preprocessing
     ├─── video2im.py     : Converts video into images
├─── data_loading.py      : Custom dataset and data processing routine
├─── main.py              : Main script to train and test the code
├─── models.py            : Number of architecture definition to try out
├─── train.py             : Methods to train, test and predict given model
├─── utils.py             : Utilities such as save_images, denorm image.

Results

TO BE UPDATED SOON