Joint detection and tracking model named DEFT, or ``Detection Embeddings for Tracking." Our approach relies on an appearance-based object matching network jointly-learned with an underlying object detection network. An LSTM is also added to capture motion constraints.
Hope you are doing well. We are trying to implement a motion prediction model different from LSTM. While trying the train_prediction.py using the MOT dataset, the code entered an endless recursive loop in the dataloader and we do not know why.
In addition, we do not understand what is the input for train_pediction.py based on train.py?
Also this command does not work to download the model file:
gdown --id 1H7RX3M7wUA55RxeZDkloby9foGgYlhcp -O src/models/model_mot.pth
In general, do you have a report or thesis that explains how did you implement the motion prediction model?
AA Muhamed,
Hope you are doing well. We are trying to implement a motion prediction model different from LSTM. While trying the train_prediction.py using the MOT dataset, the code entered an endless recursive loop in the dataloader and we do not know why.
In addition, we do not understand what is the input for train_pediction.py based on train.py?
Also this command does not work to download the model file: gdown --id 1H7RX3M7wUA55RxeZDkloby9foGgYlhcp -O src/models/model_mot.pth
In general, do you have a report or thesis that explains how did you implement the motion prediction model?
Thanks, Yasser