Hokkaido University · Computer Science
Professor Guang Li's research lab specializes in self-supervised and weakly supervised representation learning for medical image analysis, with a strong focus on addressing data scarcity and limited annotation in clinical settings. The lab develops advanced deep learning methods—particularly based on siamese networks, triplet networks, and knowledge distillation—for medical image synthesis, disease detection (e.g., COVID-19, gastritis), and model generalization under small-batch training conditions. Key research directions include robust self-supervised representation learning, dataset distillation, and efficient deep learning frameworks tailored for medical imaging applications with minimal human annotation. The lab’s work bridges the gap between theoretical advances in representation learning and practical clinical needs in radiology and diagnostic imaging.
Figures are computed from collected data and may differ slightly.
The synthesis of medical images from one modality to another is an intensity transformation between two images acquired from different medical devices, such as Magnetic Resonance (MR)image synthesis to Computed Tomography (CT)image, or MR T1 weighted (T1W)image to T2 weighted (T2W)or proton density weighted (PDW)image. MR based synthetic CT is very useful for some clinical cases, such as PET attenuation correction for PET/MR, MR/CT registration etc. In this paper, we propose a novel method based
The global outbreak of the Coronavirus 2019 (COVID-19) has overloaded worldwide healthcare systems. Computer-aided diagnosis for COVID-19 fast detection and patient triage is becoming critical. This paper proposes a novel self-knowledge distillation based self-supervised learning method for COVID-19 detection from chest X-ray images. Our method can use self-knowledge of images based on similarities of their visual features for self-supervised learning. Experimental results show that our method a
This paper proposes a novel self-supervised learning method for learning better representations with small batch sizes. Many self-supervised learning methods based on certain forms of the siamese network have emerged and received significant attention. However, these methods need to use large batch sizes to learn good representations and require heavy computational resources. We present a new triplet network combined with a triple-view loss to improve the performance of self-supervised represent
In this study, we propose a novel dataset distillation method based on parameter pruning. The proposed method can synthesize more robust distilled datasets and improve distillation performance by pruning difficult-to-match parameters during the distillation process. Experimental results on two benchmark datasets show the superiority of the proposed method.
Manually annotating gastric X-ray images for gastritis detection is time-consuming and expensive because it typically requires expert knowledge. In this paper, we propose a novel self-supervised learning method for gastritis detection with scarce annotations. Our method introduces triplet networks and a triple-view loss to solve the insufficient available annotations in gastritis detection. Experimental results show that our method can outperform several state-of-the-art methods for gastritis de
A novel cross-view self-supervised learning (CVSSL) method via momentum statistics in batch normalization is presented in this paper. The problem of accuracy degradation in small-batch cases is currently common in self-supervised learning. Our method introduces the cross-view loss and the momentum statistics in batch normalization to solve the accuracy degradation problem in small-batch cases. Experimental results show that our method can drastically outperform the state-of-the-art self-supervis
Image registration, segmentation, and visualization are three major components of medical image processing. Three-dimensional (3D) digital medical images are three dimensionally reconstructed, often with minor artifacts, and with limited spatial resolution and gray scale, unlike common digital pictures. Because of these limitations, image filtering is often performed before the images are viewed and further processed (Behrenbruch, Petroudi, Bond, et al., 2004). Different 3D imaging modalities us
Deep convolutional neural networks (DCNNs) have been popular with medical image classification problems in recent years. However, training a DCNN model on the sizeable medical dataset requires repeated manipulation to achieve the desired results and hence is time-consuming. Since there is an inevitable link between DCNN training results and the complexity of the medical dataset, it is essential to accurately evaluate the medical dataset's complexity before training the DCNN models. In this paper
In this study, we propose a novel method for algal bed region segmentation using aerial images. Accurately determining the carbon dioxide absorption capacity of coastal algae requires measurements of algal bed regions. However, conventional manual measurement methods are resource-intensive and time-consuming, which hinders the advancement of the field. To solve these problems, we propose a novel method for automatic algal bed region segmentation using aerial images. In our method, we use an adva
Open papers in the app to read, cite, and organize with AI.
