ML Kit via TensorFlow Lite

In early 2017, the deep learning framework TensorFlow was released as an open source project by  Google Brain. TensorFlow has one of the biggest and most vibrant community and  has a much bigger community behind it than PyTorch.
On November 14th , 2017, Google announced the developer preview of TensorFlow Lite for mobile and embedded devices.  TensorFlow lite provides better performance and smaller binary size than TensorFlow for Mobile which was designed to be a deep learning solution for mobile platforms. However TensorFlow lite  is in developer preview  so not all use cases are covered yet.
In collaboration with Apple, TensorFlow Lite supports CoreML through the TensorFlow Lite format (.tflite).   Google application like as hey google, Smart Reply or portrait mode on android camera are developed  using Tensor Flow lite.
Cross-platform for TensorFlow Lite

On May 8th, Google 
presented a new machine learning SDK called ML Kit in beta.  It uses TensorFlow Lite and  brings Google's machine learning expertise to mobile developers.  Actually  ML Kit provides five ready-to-use APIs :
  • Text recognition
  • Face detection
  • Barcode scanning
  • Image labeling
  • Landmark recognition

We have tested ML Kit using Android studio. The face detection API firstly locates human faces in the visual media. Secondly the application searches for landmarks such as the left eye, right eye, and base of the nose. Additionally, it provides basic emotion recognition.
Face Detection using ML Kit

We also had tested text recognition APIs which can recognize text in any Latin-Based language.

Text recognition using ML Kit


Popular Posts