Inspiration
Wanting to use computer vision and machine learning techniques to make a fun game.
What it does
The program presents you with a situation of trying to enter a night club. The bouncer asks you several questions and the program will gauge your response to the questions throurgh speech recognition and facial expression. Depending on the response to the questions, the user can either get into the club or get rejected.
How we built it
- Facial expression analysis - Detect a face in a video stream and analyze facial expression using pre-trained neural network model using Google's tensorflow.
- Speech recognition - Recognize speech and convert it to text using Google's Speechrecognition library.
- Sentiment Analysis - Analyze the sentiment of the converted text using TextBlob python library..
Challenges we ran into
OpenCV does not support audio processing. So we had to find an alternative library to do video processing.
Accomplishments that we're proud of
Getting the emotion recognition and speech recognition to work well.
What we learned
Emotion recognition using tensorflow.
What's next for Face Your Lies
Dealing with more responses and making it more interactive with the user.
Log in or sign up for Devpost to join the conversation.