CPCB student Mostofa Uddin's work accepted by CVPR
Mostofa Rafid Uddin, a second-year Ph.D. student in Dr. Min Xu’s lab, has had a first-authored paper accepted by the EEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), a top-ranked conference in computer science.
The paper is titled:
Harmony: A Generic Unsupervised Approach for Disentangling Semantic Content from Parameterized Transformations
In this paper, Mostofa and his co-authors developed an unsupervised method for disentangling semantic content from parameterized transformations, which has the potential to facilitate unsupervised detection and recovery of macromolecular structures from electron tomography data. “Though the method is primarily motivated for cryo-electron tomography image analysis, it generalizes well to many wild-type natural images. This provides a good demonstration of how two different fields of computational biology and computer vision can benefit from each other.” says Mostofa.