MOT_test_Khalilullah

Data Annotation

I have annotated one video (Test_Lacrosse_1080.mp4) using Darklabel annotation software darklabel. I shall use this video for training after completed a model preparation for tracking.

Training

data preparation from annotation data

Use data_prepa_labels_file.py and data_prepa_generate_path_file.py to create label files and path file from the annotated data. Download pretrained model, configuration file, and associated folders from this link.models

Then replace the folders in the MOT_test_Khalilullah folder. run the train.py file: python train.py

Detection and Tracking

run track.py file: python track.py

Qualitative results

results

Multi_object_tracking_vision

Datalink and References
Multi-Object Tracking Wiki: https://en.wikipedia.org/wiki/Multiple_object_tracking Multi-Object Tracking papers and Codes: https://paperswithcode.com/task/multi-object-tracking Learn OpenCV: https://learnopencv.com/multitracker-multiple-object-tracking-using-opencv-c-python/ MOT Challenge: https://motchallenge.net/