Rsight

On-Device Object Detector and Analyzer Showcase app, adaptable to various ML models.


Google's cutting-edge, end-to-end open source platform TensorFlow lets developers easily build and deploy ML powered applications.
Rsight utilizes 'TensorFlow Lite' for its object detector which is an open source deep learning framework for on-device inference.


Rsight utilizes SSD MobileNetV1 pre-trained model hosted on https://tfhub.dev/


How to use Recognizer for better performance (simple user guide)?
- To detect and save an objects simply point the camera of your smartphone
onto the sight with clear background.
- To display objects in your preferred 1 of 4 languages (English, Russian,
Spanish, Chinese) simply choose from Preferences.
- For high accurate results change the percentage of threshold confidence
level from Preferences.
- In list view you can analyze detected objects by searching, displaying the
highest confidence for each class, delete by swipe or erase all.