Human Detection from Videos based on Deep Learning

Since the state-of-the-art detection performances are mainly from still images, besides the existing algorithms validation, effective algorithms specifically designed for feature extraction and human detection from videos are expected to be developed or extended from algorithms for general object detection and human detection tasks. What’s more, it will be possible to analyze human actions like tracking customers’ paths, measuring the number of people entering or leaving the store based on the extracted features and human detection results above. “Smart Retail” project started by Sengled Canada in September, 2016 is highly related to human detection and human action understanding field.

Faculty Supervisor:

Greg Mori

Student:

Partner:

Sengled Canada

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Simon Fraser University

Program: