This is a custom image classifier which is powered by Tensorflow.js . You can train your classifier (4 classes) on the fly in the browser. This application is using pre-trained model for image classification called MobileNet. Your model will be trained on top of this one to customize the image classes it recognizes. So you can create highly accurate model with minimal training data.
- Allow your browser to reach you webcam. Your webcam will be source of your image inputs
- You can set 4 different classes at the application. Click on the related class at Training section while you are showing it to the webcam
- After the training you can show one of 4 classes to the webcam and you can see prediction and probability at the Output section.
TensorFlow.js, an open-source library you can use to define, train, and run machine learning models entirely in the browser, using JavaScript and a high-level layers API. If you’re a JavaScript developer who’s new to ML, TensorFlow.js is a great way to begin learning.
TensorFlow Lite is a set of tools to help developers run TensorFlow models on mobile, embedded, and IoT devices. It enables on-device machine learning inference with low latency and small binary size. This led to pre-trained deep learning models designed for mobile devices. There are many pre-trained models made for TensorFlow.js. We will be using MobileNet for Image Classification.
MobileNets are small, low-latency, low-power models parameterized to meet the resource constraints of a variety of use cases. They can be built upon for classification, detection, embedding and segmentation.