Scholar
Rajat Koner
Google Scholar ID: yuOQZtcAAAAJ
LMU, Munich, intern at DeepMind
Citations & Impact
Citations
466
H-index
10
i10-index
10
Publications
19
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Publications include Relationformer: A Unified Framework for Image-to-Graph Generation (ECCV, 2022), OODformer: Out-Of-Distribution Detection Transformer (arXiv, 2021), etc. Research projects have been accepted and published at multiple international conferences.
Research Experience
- University of Munich: 2018.10 - present, Doctoral Researcher, Topic: Scene Graph, Exploration of Context in an Image or Video; Fortiss: 2017.01 - 2018.04, Research Intern, Topic: Comprehensive Fore-sight of Traffic Situation through Object Detection and Tracking; J.P Morgan Chase/Cognizant: 2011.12 - 2015.09, Software Developer, Topic: a one-stop OFAC payment processing and workflow investigation.
Education
- Ph.D. Candidate: 2018.12 - present, Deep Learning, University of Munich, advised by Prof. Volker Tresp; Master Degree: 2015.10 - 2018.07, Informatics, Technical University of Munich (TUM), Thesis on Real-Time Multi-Object Tracking using Recurrent Neural Network, supervised by Prof. Alois Knoll; Bachelor Degree: 2007.08 - 2011.08, Computer Science Engineering, West Bengal University of Technology, Thesis on Object Detection for Embedded System, supervised by Prof. Amlan Chakrabarti.
Background
- Research interests include computer vision and deep learning, especially in Scene Graph and Scene Understanding, Visual Question Answering using Scene Graph Reasoning, Image and Object Understanding with Cognitive Models (e.g., Transformers), Semi/Self-supervised Representation Learning for Outlier Detection.
Co-authors