The TensorFlow.js coco-ssd (Common Objects in Context Single Shot Detection) model takes input from the user's camera, and returns an array of bounding boxes with class name and confidence level which are displayed using React components.
The model can detect 80 classes of objects. Here is a full list of the object classes.
- HTML, CSS, JavaScript
- TensorFlow.js
- React.js
- Node.JS
- Allow the user to upload files for detection.
Object Detection model (coco-ssd) from pre-trained TensorFlow.js library.
More info on the COCO dataset.
Using hooks with ref callbacks