Published in The 23rd IEEE International Conference on Bioinformatics and Bioengineering (BIBE), 2023
Developed a novel self-supervised pipeline estimates an Out-of-Distribution score for each unlabeled data point, enabling calibrated semi-supervised learning and improving medical image analysis, with potential applications in open-set supervised learning. Read more
Published in 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), 2022
Developed an attention-based deep learning classifier framework that effectively integrates both clinical and imaging features to automate the analysis of renal tumors, leading to a significant 12.5% improvement in classification accuracy. Read more
Published in 2022 IEEE 18th International Symposium on Biomedical Imaging (ISBI), 2022
Developed a deep learning classification framework leveraging contrastive learning-based clustering and memory banks to prioritize clean samples, achieving a 13.4% improvement in accuracy on medical images with both closed-set and open-set label noise. Read more