Use mmtracking for text tracking

open-mmlab / mmtracking

OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.

https://mmtracking.readthedocs.io/en/latest/

Apache License 2.0

3.53k stars 591 forks source link

Use mmtracking for text tracking #502

Open anhquan075 opened 2 years ago

anhquan075 commented 2 years ago

Hello MMTracking team, I want to use some datasets in video text spots to train for the video object detection task. Is it possible to do this?

Seerkfang commented 2 years ago

Yes. But you may modify the size of anchors according to the traits of the text.