Bringing ease to Blind Individuals interaction with objects in their day to day lifestyle by creating an application powered by Machine Learning and Neural Networks
TensorFlow and Keras Libraries was utilized to develop the application applying Transfer Learning with Mobile Nets to create the Model. The generated TF Lite Model was used on the Android Device. The Android Application will use to recognize images.
![](images/blog/app1.1.jpg)
- Pretrained and Optimized Customizable Image Classification Model using Images captured from the user's homespace to Identify The Objects in the House.
- Mobile Nets V2 – State of the art mobile architecture processing images on Mobile & IOT Devices in Realtime using Low Memory Footprint
- Uses Depth wise Separable Convolutional Neural Networks, Linear Bottlenecks between the layers & shortcut connectors between the bottlenecks (Residual Connections). .
For more info check