Prakash Chandra Chhipa
Doctoral Researcher, Luleå Universtiy of Technology, Sweden
A35374, Luleå Universtiy of Technology, Sweden
Currently, Prakash Chandra Chhipa is a dedicated Computer Vision Researcher and Ph.D. Candidate at the Machine Learning Group at Luleå University of Technology, Sweden, advised by Prof. Marcus Liwicki.
Before embarking on his academic journey, Prakash contributed significantly to industrial R&D for over eight years, specializing in machine learning and computer vision with several multinational companies. His experience spans identifying challenges, cultivating research ideas, developing prototypes, inventing intellectual properties, and transforming them into commercialized products.
Prakash’s research is deeply rooted in computer vision and representation learning, with a particular focus on self-supervised learning (SSL). His doctoral work is two-fold: first, he focuses on robustness in self-supervised representation learning to handle out-of-distribution (OOD) performance, focusing on mitigating the impact of natural and synthetic distribution shifts in real-world scenarios. Second, he aims to develop domain-aware SSL methods tailored to diverse visual domains, such as medical imaging, remote sensing, and RGB-D images, where visual concepts significantly deviate from natural scenes. Prakash’s journey is driven by a genuine desire to bridge the gap between innovative research and impactful real-world solutions.
For more details, please check his updated CV here.
Academic Research Highlights
PhD candidate at ML Group, Luleå University of Technology, Sweden. Defended my licentiate thesis with emphasis on adaptation of self-supervised representation learning methods on beyond natural scene visual domain in March 2023 and now focusing on robust representation learning towards my doctoral thesis. Published rsearch works in top conferences including ECCV, WACV, ICIP and ICCV/ECCV workshops. Have been a reviewer at CVPR, NeurIPS, and other conferences, industrial research with Swedish companies, and secured multiple grants.
Visiting Researcher at Center for Research in Computer Vision (CRCV), University of Central Florida, USA. Guided by host Prof. Mubarak Shah. Working on robust representation learning and also focusing Geo-localization domain. (Mar-June 2024)
Invited talk at EMBL-European Bioinformatics Institute at Cambridge, UK and MIRAI 2.0 Research and Innovation week at Kyushu University, Fukuoka, Japan. Accepted and participated in machine learning school at Cambridge University, UK and other venues.
Industrial R&D Work Tenure
Built a team and led computer vision research and development in the microscopic histopathology visual domain at Arkray R&D. Proudly contributed to achieving novel sediment recognition method on low-resolution histopathology for successful commercialization of Urinary Sediment Analyzer AUTION EYE AI-4510. Have been inventor for multiple patent applications (2018-2020).
Machine learning at AI & Data Intelligence Group (Now Advanced Research - Vision AI), Samsung R&D Institute for 6+ years (2012-2018) in different roles. Have been in forefornt for long-term R&D of successful large-scale products Samsung ACR and Samsung Ads focusing majorly on computer vision and reinforcement learning. Have been inventor for multiple granted patents/IP applications and awardee for Innovator of the Year-2017 for multiple IPs.
Worked with HCL Technologies for two years (2010-2012) and have been a contributor for enterprise-scale development of a platform for leading Telcom service provider T-Mobile.
news
Nov 02, 2024 | 🥇 Featured in Swedish Trans-Atlantic Researchers and Scholars Network (STARS) for recent work. 🥇 |
---|---|
Oct 29, 2024 | 🎓 Accepted as DAAD AI-Net Fellow 2024 on subject ‘AI for Science’ with networking tour to host German institute for potential collaboration in Artificial Intelligence. 🎓 |
Sep 20, 2024 | 🎉 Paper titled ‘LCM: Log Conformal Maps for Robust Representation Learning to Mitigate Perspective Distortion’ accepted in ACCV 2024. 🎉 |
Aug 15, 2024 | 🎉 Paper titled ‘Open-Vocabulary Object Detectors: Robustness Challenges under Distribution Shifts’ accepted in ECCV OOD-CV workshop 2024. 🎉 |
Jul 03, 2024 | 🎉 Paper titled ‘Möbius Transform for Mitigating Perspective Distortions in Representation Learning’ accepted in ECCV 2024. 🎉 |