https://arxiv.org/pdf/1608.06197.pdf
I have read quite a few articles and papers surrounding this topic. However, most of my understanding comes from analysing the paper called “Dilated Convolutional neural network for Understanding the Highly congested scene”. Here are some notes I took on the iPad along with the paper:
After fully understanding this paper, other articles were understood more quickly, if you are interested in reading more papers on this topic, please visit this git repo:
https://github.com/gjy3035/Awesome-Crowd-Counting
This is a git repo compiled by several contributors around the world; they evaluate the papers and update them in the git repo. The ones I have read are in google drive, and I have also looked at their code to see how different ones work.
There are many paper so as a form of notes, I take screenshots of their proposed model.
Research Dump:
CrowdNet: A Deep Convolutional Network for Dense Crowd Counting
https://arxiv.org/pdf/1608.06197.pdf
To Choose or to Fuse? Scale Selection for Crowd Counting
https://ojs.aaai.org/index.php/AAAI/article/view/16360
This paper was really hard to understand; this was MCNN decoded-based network. Something I won’t look into.
Cross-scene Crowd Counting via Deep Convolutional Neural Networks