We are building a lie detection system that combines audio and video data of a subject to detect instances of deception. The idea is based on proven psycological research that shows that when an individual lies, there are changes in his/her voice (voice amplitude, energy, pitch, vowel duration etc.) and facial patterns (blink patterns, micro-expressions, gaze direction etc.) that can be tracked using software.
Check out our Web Application: In Progress
In the Video below, Cliff Lansley, Director of Emotional Intelligence Academy explains exactly what changes occur in a person's voice and facial patterns when he/she lies
-
Our deliverable will be a web application, that is being managed in a separate repo here
-
All of our presentations and research material can be found here
-
Our tasks are organized on trello. You can check and see our progress here
-
Huffman: Implementation of Huffman Algorithm and its Algorithmic Proof
-
Searcher: Comparision of Google vs. Bing using Selenium webdriver to automate searches to determine which one is better at autocorrecting a mispelled word and providing its definition
-
ThinkDSP: Forked Project: A tool for voice analysis
-
Voice Detection: Utility functions for detecting voice components such as vowel duration, pitch etc.
to be used in final web application -
Blink Detection: Analyzes video to determine when a blink occurs.
to be used in final web application