Yue Yao

I am currently a Research Officer at the School of Computing of the Australian National University, working with Prof. Tom Gedeon. Before this role, I completed my PhD at ANU, during which I was fortunate to get supervision from Prof. Tom Gedeon and Prof. Liang Zheng.

During my PhD, I interned at NVIDIA, working on Metropolis with Dr. Milind Naphade and Dr. Thomas Tang.

Email / Google Scholar / Github

News

I am co-organzing 3rd Workshop on Vision Datasets Understanding and 2nd DataCV challenge in CVPR 2024.

I am co-organizing the 2024 AI city challenge in CVPR 2024.

Attribute Descent: Simulating Object-Centric Datasets on the Content Level and Beyond, TPAMI, Accepted. [PDF] [bibtex]

Research

I'm interested in computer vision, deep learning, and brain computer interface. Much of my research is about optimizing the training set to improve model specificity on target domains, which is also my PhD topic. Selected publications are listed below.

	Alice Benchmarks: Connecting Real World Object Re-Identification with the Synthetic Xiaoxiao Sun, Yue Yao, Shengjin Wang, Hongdong Li, Liang Zheng ICLR, 2024 PDF Within the Alice benchmarks, two object re-ID tasks are offered: person and vehicle re-ID. We collected and annotated two challenging real-world target datasets: AlicePerson and AliceVehicle for Syn2real re-ID research.
	Attribute Descent: Simulating Object-Centric Datasets on the Content Level and Beyond Yue Yao, Liang Zheng, Xiaodong Yang, Milind Napthade, Tom Gedeon TPAMI, 2024 PDF / Code We aim to use graphic engines to simulate a large number of training data that have free annotations and possibly strongly resemble to real-world data.
	Open-Set Facial Expression Recognition Yuhang Zhang, Yue Yao, Xuannan Liu, Lixiong Qin, Wenjing Wang, Weihong Deng AAAI, 2024 PDF We propose the open-set facial expression recognition task, and an strong baseline model for this task.
	Optimizing the Training Set to Improve Model Specificity on Target Domains Yue Yao PhD Thesis, 2023
	Training with Product Digital Twins for AutoRetail Checkout Yue Yao, Xinyu Tian, Zheng Tang, Sujit Biswas, Huan Lei, Tom Gedeon, Liang Zheng (* denotes equal contribution) Arxiv, 2023 PDF / Code We aim to address the challenge in the absence of a labeled real-world training set for AutoRetail Checkout. We address this by data rendering from the given 3D retail models.
	Large-scale Training Data Search for Object Re-identification Yue Yao, Huan Lei, Tom Gedeon, Liang Zheng CVPR, 2023 PDF / Code We present a search and pruning (SnP) solution to the training data search problem in object re-ID. It results in a training set 80% smaller than the source pool while achieving a similar or even higher re-ID accuracy.
	Information-preserving Feature Filter for Short-term EEG Signals Yue Yao, Josephine Plested, Tom Gedeon Neurocomputing, 2020 Paper / Code We studied the feature fusion problem in EEG, and proposed a feature filter to filter out unwanted (e.g., privacy-related) features from EEG data.
	Simulating Content Consistent Vehicle Datasets with Attribute Descent Yue Yao, Liang Zheng, Xiaodong Yang, Milind Napthade, Tom Gedeon ECCV, 2020 Paper / Code / Demo / Video We used content-adapted synthetic vehicle data to augment real vehicle re-ID data and get improved accuracy.

The source code of this website comes from Jon Barron's website