About Me

I am a Ph.D. student affiliated with The Vision Lab at IIIT Delhi, India, working with Dr. Saket Anand. My academic pursuits revolve around the dynamic field of Machine Learning and Deep Learning, particularly their applications to address intricate real-world challenges. My research is centered on computer vision, with a primary focus on robust feature representation, uncertainty estimation, and dynamic system modeling as applied to fundamental tasks such as object recognition and tracking. In addition to my core expertise, I bring interdisciplinary skills spanning theoretical subjects like 3D computer vision, graphs, differential geometry, and reproducing kernel Hilbert spaces, as well as applied areas including domain adaptation, active learning, autonomous driving, and remote sensing.

Currently, I am working as a Research Fellow on “Graph-Based Statistical Analysis of Entire Scenes by Combining Multi-Sensor, Multi-Perspective Video Streams” in collaboration with Dr. Anuj Srivastava (Department of Statistics, Florida State University). The project aims to perform a spatiotemporal fusion of multiple asynchronous dynamic scenes captured from multi-modal sensors using a sensor-agnostic data representation (scene graphs). In the past, I was also supported by Google AI for Social Good Program for the work in collaboration with Dr. T. Jayaraman (MSSRF) on analyzing the impact of aridification on agricultural production in the Cauvery Delta using multi-modal cross-satellite data. An outcome of this work is a first-of-its-kind dataset, SICKLE (Satellite Imagery for Cropping annotated with Key-parameter LabEls), accepted at WACV'24 as an oral presentation. We were also recognized for our methodology in the Google AI4SG Mid Program Workshop. I also got to present one of our works at ECCV'22 on “Learning Hierarchy Aware Features for Reducing Mistake Severity”, a collaborative effort with my colleagues Ashima Garg and Dr. Saket Anand. The primary motivation behind the problem is to reduce the severity of mistakes while not compromising on the top-1 accuracy by incorporating prior knowledge about the hierarchy of the data labels.

Before joining the Ph.D. program, I worked as a Software Research Engineer at Samsung Research Institute in the Software Innovation & Intelligence Department. My responsibility included the Research and Development of innovative inventions for the organization. Some of my work at the organization included an On-Device Dynamic Emoji Generation solution that merges two or more emojis to create a new one, a Face-Expression & Movement Based Pattern Matching Authentication Mechanism that authenticates the user not just based on the facial pattern but also recognizes the face-expression pattern set by the user. 

Recent Publications

Awards & Recognitions

• Contest Organizer, Anveshan 2.0: Organizing a LiDAR-based drivable region segmentation challenge for Anveshan 2.0

• Ph.D. Mentor, CoDE-AI: Serving as a mentor for Competitive Data Engineering and Artificial Intelligence (CoDE-AI), which is a student group that participates and competes in data challenges.

• Organizing Committee Member, Pasteur’s Quadrant Seminar Series: PQSS is a student-led initiative in the AI for social good space that aims to increase awareness and build a community of students and practitioners who are excited to contribute to this field. The aim of this program is also to foster an AI for social good ecosystem that brings together experts from various disciplines and establishes AI4SG as an interdisciplinary science

• Contest Organizer, ICVGIP Contest 2021: Organized an AI contest for the Indian Conference on Computer Vision, Graphics and Image Processing.

Get in touch at depanshus@iiitd.ac.in