boxmot

The project can achieve FCWS, LDWS, and LKAS functions solely using only visual sensors. using YOLOv5 / YOLOv5-lite / YOLOv6 / YOLOv7 / YOLOv8 / YOLOv9 / EfficientDet and Ultra-Fast-Lane-Detection-v2 .

基于TensorRT的C++高性能推理库，Yolov10, YoloPv2，Yolov5/7/X/8，RT-DETR，单目标跟踪OSTrack、LightTrack。

?? CVPR 2025 (Nashville TN) - The 4th Anti-UAV Workshop & Challenge ?

Анализ трафика на круговом движении с использованием компьютерного зрения

Integration of YOLOv9 with ByteTracker

Tracking-by-Detection形式のMOT(Multi Object Tracking)について、 DetectionとTrackingの処理を分離して寄せ集めたフレームワーク(Tracking-by-Detection method MOT(Multi Object Tracking) is a framework that separates the processing of Detection and Tracking.)

juxtapose: Multi-Person Pose Tracking Inference SDK with RTMDet, YOLOv8, GDino, RTMPose (ONNX) & Trackers (ByteTrack & BotSORT) & Tapnet with custom ROIs + FastAPI GPU exe

ByteTrack(Multi-Object Tracking by Associating Every Detection Box)のPythonでのONNX推論サンプル

yolox-bytetrack-sampleのマルチクラス拡張版

YOLOXとByteTrackを用いたMOT(Multiple Object Tracking)のPythonサンプル