An invited talk by Prof. Yu Wang, Associate Professor, Hitotsubashi University, Japan, will be held on December 7, 2022, at Zoom (online). Prof. Yu Wang will share some interesting ideas about Pedestrian Detection by Learning Shallow Classifiers on Deep Features.

Date: 07 Dec 2022
Time: 08:45 AM to 10:15 AM (JST)

Location:

https://zoom.us/j/7701243471?pwd=ZFcrcmNPbmxtRC9qbTlackc3bWdtZz09

CO-ORGANIZED BY:

IEEE Muroran Institute of Technology Student Branch

IEEE Sapporo Section Young Professionals (YP)

Emerging Networks and Systems Laboratory (ENeS), Muroran Institute of Technology

Topic: Pedestrian Detection by Learning Shallow Classifiers on Deep Features

Abstract:

This talk briefly introduces two works for pedestrian detection in which deep features are coupled with classical shallow classifiers. The first work adopts body parts mining on deep features. By learning and transforming human body part detectors to DCNN layers, we proposed the DP-CNN. The second work adopts a large-scale performance evaluation of the local deep features. Based on the results, we proposed the MS-CNN. By evaluating the above two deep models, we show that the shallow learning approaches not only work well with the deep representations, but also motivate novel end-to-end deep model architectures that are powerful.