For a complete list of my publications, please refer to my Google Scholar page.

My current research is on the analysis of fluorescent microscopy images. I have two main directions in my current research pursuit: (i) Developing better segmentation algorithms for biomedical data, particularly for time-lapse microscopy images (ii) Accelerating compilation of new biomedical datasets.

Better Segmentation Algorithms for Biomedical Data

alt text 

One crucial component of quantitative analysis of biomedical data is to develop a robust segmentation solution. While state-of-the-art segmentation algorithms are applicable to this problem, there are challenges and opportunities specific to this type of data. Our goal in this project is to develop segmentation solutions for the analysis of biomedical datasets. In particular, we are interested in analyzing time-lapse microscopy images. Our approach is centered around exploiting multi modal nature of the data. In addition, we treat the sequences of time-lapse microscopy images as video and we attempt to model temporal dependencies across the frames using deep learning architectures. We have published a workshop paper in CVPR2017 on this topic.

Efficient Biomedical Dataset Compilation

Compiling biomedical datasets is often necessary to train or fine-tune existing models for segmentation. However, this process is often costly for researchers, usually on the order of hours per manuscript. In this project, we attempt to mitigate this problem by (i) developing annotation interfaces for dataset creation (ii) devising smart data augmentation techniques.