A task that involves real-time object detection of person objects, conversion into a specified input format, and then feeding into ViT (Vision Transformer) as input:
- Detect person objects with bounding box detection.
- Track the detected person objects in real-time.
- Listen in real-time, and simultaneously, have the ViT also listen.
| Inference | Def |
|---|---|
| hog_test.py | Setup of HOG Descriptor and Default person Detector |
| webcam_test.py | Setup of person Object Detection Using Built-in Webcam |
| yolo_main.py | Setup of person Object Detection Using YOLO_NAS, Tracking TOOL Deepsort |
| Utils | Def |
|---|---|
| image_preprocessing.py | Crop the detected person object, save it in .jpg format, and convert it into a vector suitable for input into a Vision Transformer |
| tracking.py | Utilize YOLO for detecting person objects and employ DeepSort for real-time tracking |
.
├── README.md
├── args_parser.py
├── deep_sort_pytorch
│ ├── configs
│ │ └── deep_sort.yaml
│ ├── deep_sort
│ │ ├── README.md
│ │ ├── deep
│ │ ├── deep_sort.py
│ │ ├── sort
│ │ └── sort - Copy
│ └── utils
│ ├── asserts.py
│ ├── draw.py
│ ├── evaluation.py
│ ├── io.py
│ ├── json_logger.py
│ ├── log.py
│ ├── parser.py
│ └── tools.py
├── input
├── output
├── ultralytics
│ ├── docker
│ ├── docs
│ ├── examples
│ ├── mkdocs.yml
│ ├── pyproject.toml
│ ├── tests
│ └── ultralytics
├── utils
│ ├── image_preprocess.py
│ └── tracking.py
├── webcam_test.py
├── yolo_main.py
├── yolo_nas.py
├── yolo_pt
├── yolo_save_csv
└── yolo_save_img