Structured Gradient-based Interpretations via Norm-Regularized Adversarial Training

1 code implementation • 6 Apr 2024 • Shizhan Gong, Qi Dou, Farzan Farnia

We show a duality relation between the regularized norms of the adversarial perturbations and gradient-based maps, based on which we design adversarial training loss functions promoting sparsity and group-sparsity properties in simple gradient maps.

3DSAM-adapter: Holistic Adaptation of SAM from 2D to 3D for Promptable Medical Image Segmentation

1 code implementation • 23 Jun 2023 • Shizhan Gong, Yuan Zhong, Wenao Ma, Jinpeng Li, Zhao Wang, Jingyang Zhang, Pheng-Ann Heng, Qi Dou

Notably, the original SAM architecture is designed for 2D natural images, therefore would not be able to extract the 3D spatial information from volumetric medical data effectively.

Image Segmentation Medical Image Segmentation +2

Diffusion Model based Semi-supervised Learning on Brain Hemorrhage Images for Efficient Midline Shift Quantification

no code implementations • 1 Jan 2023 • Shizhan Gong, Cheng Chen, Yuqi Gong, Nga Yan Chan, Wenao Ma, Calvin Hoi-Kwan Mak, Jill Abrigo, Qi Dou

Brain midline shift (MLS) is one of the most critical factors to be considered for clinical diagnosis and treatment decision-making for intracranial hemorrhage.

Decision Making Representation Learning