SCALE@NTU Research Webinar Jan 2022

This research webinar on Teams is organized by Singtel Cognitive and Artificial Intelligence Lab for Enterprises (SCALE@NTU) to share the research work in the Corp Lab. For registration, please visit:

https://wis.ntu.edu.sg/pls/webexe88/REGISTER_NTU.REGISTER?EVENT_ID=OA22010617103580

Talk 1:  Low Resolution Also Matters: Learning Multi-Resolution Representations for Person Re-Identification

As a prevailing task in video surveillance and forensics field, person re-identification (re-ID) aims to match person images captured from non-overlapped cameras. In unconstrained scenarios, person images often suffer from the resolution mismatch problem, i.e., Cross-Resolution Person Re-ID. To overcome this problem, most existing methods restore low resolution (LR) images to high resolution (HR) by super-resolution (SR). However, they only focus on the HR feature extraction and ignore the valid information from original LR images. In this work, we explore the influence of resolutions on feature extraction and develop a novel method for cross-resolution person re-ID called Multi-Resolution Representations Joint Learning (MRJL). Our method consists of a Resolution Reconstruction Network (RRN) and a Dual Feature Fusion Network (DFFN). The RRN uses an input image to construct a HR version and a LR version with an encoder and two decoders, while the DFFN adopts a dual-branch structure to generate person representations from multi-resolution images. Comprehensive experiments on five benchmarks verify the superiority of the proposed MRJL over the relevant state-of-the-art methods.

Speaker:  Dr Zhang Guoqing, Research Fellow, SCALE@NTU

Guoqing Zhang received his B.S. and Master degrees from Yangzhou University in 2009 and 2012, and his Ph.D. degree from Nanjing University of Science and Technology in 2017. He has been a Research Fellow at Singtel Cognitive and Artificial Intelligence Lab for Enterprises (SCALE@NTU) since Dec 2018. His research interests include computer vision, person re-identification and domain adaptive learning.

Talk 2:  Cross-View Regularization for Domain Adaptive Panoptic Segmentation

Panoptic segmentation unifies semantic segmentation and instance segmentation, which has been attracting increasing attention in recent years. However, most existing research was conducted under a supervised learning setup whereas unsupervised domain adaptive panoptic segmentation which is critical in different tasks and applications is largely neglected. We design a domain adaptive panoptic segmentation network that exploits inter-style consistency and inter-task regularization for optimal domain adaptive panoptic segmentation. The inter-style consistency leverages geometric invariance across the same image of the different styles, which ‘ fabricates’ certain self-supervisions to guide the network to learn domain-invariant features. The inter-task regularization exploits the complementary nature of instance segmentation and semantic segmentation and uses it as a constraint for better feature alignment across domains. Extensive experiments over multiple domain adaptive panoptic segmentation tasks (e.g. synthetic-to-real and real-to-real) show that our proposed network achieves superior segmentation performance as compared with the state-of-the-art.

Speaker:  Huang Jiaxing, Research Associate, SCALE@NTU

Jiaxing Huang received his B.Eng. and M.Sc. in EEE from the University of Glasgow, UK, and Nanyang Technological University (NTU), Singapore, respectively. He is currently a Research Associate in SCALE@NTU and a Ph.D. student with School of Computer Science and Engineering, NTU, Singapore. His research interests include computer vision and machine learning.