Projects
Visual Sentiment Analysis
Analysed different models such as LSTM, CNN, DAN to extract the sentiment of any image using the image captioning text from show and tell model.
Search and rescue drone with autonomous exploration
Built an autonomous navigation model to navigate drones inside buildings and perform object detection to assist first responders with search and rescue activities.
Sign2Speak
Trained and built a CNN model based on inception-v3, integrated with an android application which converts Indian sign language to English text.
Speech Assisted Interface Library
SAIL (Speech Assisted Interface Library) is a library that will make your application much more accessible, in the most simplest way you can think of... speech. You can let your users talk to the application as if it were an actual person, thanks to the beauty of Machine Learning, and let a whole variety of new users to literally converse with your application.
Pothole Detector
Android application to detect pothole on roads. App runs as a background service and is activated only when the app is in driving mode. Makes use of the accelerometer and gyroscope of the phone to detect when the vehicle goes over a pothole. Once a pothole is detected, it uses location services to update a database with the pothole locations. On the user interface of the app, using google map API, we suggest the fastest route with least number of potholes.