Deepu Raveendran

Computer Vision & Deep Learning Researcher, Samsung R&D deepu.raveendran71@gmail.com

I am a researcher in computer vision and deep learning. I have more than 10 years of experiance in Deep Learning and Machine Learning models. I am currently working as Senior Chief Engineer at Samsung R&D. My current research area is focussed on developing accurate video instance segmentation and high quality image matting.

I have strong mathematical foundations on generative models such as VAE, GAN and Diffusion models. Currently my research area is focussed on intersection of large vision based generative models and multimodal machine learning. My research focus extends to Explainable AI, Trustworthy AI, Uncertainity Estimation and Domain Adaptation. I have hands on experaince in Continual AI technique such as LoRA (Low Rank Adaptation), few shot learning, meta learning etc. In the past I have worked image similarity estimation using state of the art metric learning technique such as Multi Head Siamese network.


Experience

Senior Chief Engineer

Samsung Research Institute Bangalore

Developed portrait model for flagship phone of Samsung. I have developed selfie portrait for Samsung S22,S23 and S24. I have also created light weight instance segmentation model for video and camera preview.

I have also involved in development person reidentification for temporal consistent video instance segmenentation. I am also involved in other classical research work such as Super Resolution, Domain Adaptation and Continual AI for large vision models. I have hands on knowledge on VAE, GAN and Diffusion models.

September 2022 - Present

Research Analyst

Toshiba R&D Division

Developing multiple research solution for Image similarity problem, object detection and semantic segmentation problem. I have involved in the development of state of the art metric leraning system using multi head siamese architecture.

I have worked on classical research problem statements such Domain Adaptation Problem where I have extensively investigated generative adversarial network (GAN)-based approaches for identifying similarities between images from different domains.

September 2019 - March 2022

Senior Research Engineer

Toshiba R&D Division

Developed noval object detection framework using Yolo v3/v5, Faster RCNN, CentreNet and semantic segmentation using UNet and DeepLab. I have worked on Continual AI problem such as class incremental learning and online learning without much catastrophic forgetting on old data

September 2016 - March 2019

Research Engineer

Mathworks

I am involved in development of anisotropic diffusion for fog removal and low-light image enhancement and developed a novel technique for multi-frame super resolution. My primary role involved integrating the Point Cloud Library and SLAM (Simultaneous Localization and Mapping) into the Computer Vision Toolbox.

September 2013 - September 2016

Education

VIT University, Vellore

M.Tech
Computer Science Engineering
August 2011 - June 2013

Cochin University of Science & Technology

B.Tech
Information Technology
July 2006 - May 2010

Skills

Programming Languages & Tools
Libraries

Interests

Apart from being a Researcher, I am interested in computational photography and competeive coding. I am interested in processing remote sensed data, and astro images.

When forced indoors, I follow a number of sci-fi and fantasy genre movies and television shows, I am an aspiring chef, and I spend a large amount of my free time exploring the latest technology advancements especially in the generative imaging field.


Awards & Certifications