This project refines YOLOv8 for traffic sign detection and recognition(TSDR) with attention modules.
Base model: YOLOv8 Extra Large (YOLOv8x)
Attention modules: SE and ECA
Dataset: Tsinghua-Tencent 100K and SUSTech custom test set.
Platform: 4 * GeForce RTX 2080 Ti, Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz, SUSTech GPULab02 of CSE.
Project for SUSTech Course CS329 Machine Learning(H).
For experiments and results, please see our research report.
Traffic_Sign_Detection_Recognition
├─custom_testset
│ ├─images
│ └─videos
├─doc
│ └─img
├─models
│ ├─YOLOv8n
│ │ └─weights
│ ├─YOLOv8x
│ │ └─weights
│ ├─YOLOv8x_ECA
│ │ └─weights
│ └─YOLOv8x_SE
│ └─weights
├─src
│ ├─preprocesser
│ └─ultralytics
└─tt100k_2021
├── images
│ ├── other
│ ├── test
│ ├── train
│ └── val
├── labels
│ ├── other
│ ├── test
│ ├── train
│ └── val
└── marks