search

Kroery / DiffMOT

code for CVPR2024 paper: DiffMOT: A Real-time Diffusion-based Multiple Object Tracker with Non-linear Prediction
MIT License
379 stars 42 forks source link

what should I put for the config option "det_dir", I go thorugh the code and I have no idea what it is trying to do #5

Open Kayzwer opened 5 months ago

Kroery commented 5 months ago

Our work focuses on addressing non-linear motion in MOT. Therefore, we directly utilize the detection results from YOLOX to avoid wasting time by rerunning detection for each inference.

The "det_dir" here refers to the path containing the detection results, formatted as shown in here.

258863 commented 2 months ago

Our work focuses on addressing non-linear motion in MOT. Therefore, we directly utilize the detection results from YOLOX to avoid wasting time by rerunning detection for each inference.

The "det_dir" here refers to the path containing the detection results, formatted as shown in here.

请问reid部分的目录也是先训练模型，再生成ReID embeddings吗？ Hello, may I ask if the reid part is to train the model and convert the inference results into embeddings (ReID embeddings)?