This project utilizes YOLOv8 for the detection of laptop and mouse objects.
There are 2 classes: mouse and computer.
Roboflow was used for the labeling process. A total of 93 images were used, consisting of 69 for training, 16 for validation, and 8 for testing. The labeled images were stretched to 640x640 resolution. No augmentation was utilized. You can access the labeled data from here.
pip install ultralytics
- To get started, create a new folder in your Drive account and upload the dataset, data.yaml, and yolov8.ipynb files into it.
- In the data.yaml file, replace the train, val, and test paths with the file paths in your own Drive account.
- Finally, open the yolov8.ipynb file and run it to begin the process.
The graph below illustrates the values for mAP, loss, precision, and recall.
Run the following command:
!yolo task=detect mode=predict model=runs/detect/train/weights/best.pt conf=0.5 source=laptop_mouse_dataset/test/videos/z.mp4 save=True
Here is the final outcome:
