Recent Advances on Single Image Crowd Counting

PhD Qualifying Examination


Title: "Recent Advances on Single Image Crowd Counting"

by

Miss Haoyue BAI


Abstract:

Single image crowd counting is an important class of computer vision techniques 
for a wide range of intelligent surveillance applications. Crowd counting via 
density maps has been attracting much attention recently with the development 
of deep learning techniques.

This survey provides a comprehensive summary of recent advances for CNN-based 
single image crowd counting via density maps on three major categories: deep 
neural network designs, optimization strategies, and data processing methods. 
We also record some widely used benchmark datasets and evaluation metrics. We 
compare and analyze existing methods on the public datasets. Finally, we draw 
conclusions and discuss some future directions.


Date:			Monday, 21 December 2020

Time:                  	3:00pm - 5:00pm

Zoom meeting: 
https://zoom.us/j/92931608738?pwd=RUNrL21DVWgzYVJ6eVdnbEhVWDVLQT09

Committee Members:	Prof. Gary Chan (Supervisor)
 			Prof. Andrew Horner (Chairperson)
 			Prof. Cunsheng Ding
 			Dr. Xiaojuan Ma


**** ALL are Welcome ****