Arulkumar S

"The secret of success is this: there is no secret to success"


Resume

Recent updates

Research Interests

Person Detection, Tracking & Re-Identification

In the field of Computer Vision, my interest mainly revolves around detecting, tracking and re-identifying humans (as in surveillance).

Action Recogniton, Detection & Summarization

Recognizing actions of Humans is a challenging task and it is the need of the hour due to recent outburst of videos in social medias such as Youtube, Facebook etc.,

Multimodal Learning (Image, NLP, Audio)

It is particularly interesting for me to understand, how we, as humans, are using Multimodal knowledge from the world and in the same way, will we be able to make a machine to learn using multimodal data?

Publications

Journal articles

04 August, 2022: "Co-segmentation inspired attention module for video-based computer vision tasks" accepted in Computer Vision and Image Understanding (CVIU). [paper]

Authors: Arulkumar Subramaniam, Jayesh, Vaidya, MAM Ameen, Athira Nambiar, Anurag Mittal

Conferences

05 April, 2022: "Non-linear Motion Estimation for Video Frame Interpolation using Space-time Convolutions" accepted in CVPRW-2022: Workshop on Learned Image Compression (CLIC). [pdf][poster][code]

Authors: Saikat Dutta, Arulkumar Subramaniam, Anurag Mittal

09 October, 2021: "Co-Segmentation Aided Two-Stream Architecture for Video Captioning" accepted in WACV-2022 [pdf][poster][ppt][video]

Authors: Jayesh Vaidya, Arulkumar Subramaniam, Anurag Mittal

03 August, 2020: "Self-Attention based Feature Extractors for 3D Object Detection in Point Clouds" accepted in Workshop on Perception for Autonomous Driving, ECCV-2020 [pdf][ppt][video]

Authors: Arulkumar Subramaniam, Ashish Vaswani, Niki Parmar

20 August, 2019: "Feature Ensemble Networks with Re-ranking for Recognizing Disguised Faces in the Wild" accepted in ICCVW-2019 (Disguised Faces in the Wild)[pdf][ppt]

Authors: Arulkumar Subramaniam*, Ajay Narayanan*, Anurag Mittal

27 July, 2019: "Co-segmentation Inspired Attention Networks for Video-based Person Re-identification" accepted in ICCV-2019[pdf][code][poster]

Authors: Arulkumar Subramaniam, Athira Nambiar, Anurag Mittal

23 January, 2018: "NCC-Net: Normalized Cross Correlation Based Deep Matcher with Robustness to Illumination Variations" accepted in WACV-2018[pdf][code][video][poster]

Authors: Arulkumar Subramaniam*, Prashanth Balasubramanian*, Anurag Mittal

23 January, 2018: "A Generative Approach to Zero-Shot and Few-Shot Action Recognition" accepted in WACV-2018 [pdf][video]

Authors: Ashish Mishra, Vinay Kumar Verma, M Shiva Krishna Reddy, Arulkumar Subramaniam, Piyush Rai, Anurag Mittal

16 August, 2016: "Deep Neural Networks with Inexact matching for Person Re-Identification" accepted in NIPS-2016 [pdf][code][video][poster]

Authors: Arulkumar Subramaniam, Moitreya Chatterjee, Anurag Mittal

12 August, 2016: "Bi-modal First Impressions Recognition using Temporally Ordered Deep Audio and StochasticVisual Features" accepted in ECCV-2016 workshop on Apparent Personality Analysis. [pdf][code][ppt][gallery]

Authors: Arulkumar Subramaniam, Vismay Patel, Ashish Mishra, Prashanth Balasubramanian, Anurag Mittal

Work Experience

December-2019 to Feb-2020

Research Intern in Google Brain (Mentors: Niki parmar & Ashish Vaswani)

Topic: Self-Attention based Feature Extractors for 3D Object Detection in Point Clouds. [pdf][ppt][video]

May-2010 to July-2015

Senior Software Engineer in "Robert Bosch Engineering and Business Solutions Ltd (RBEI)"

Automotive Domain (Passive safety: Airbags), Driver development (VC++, Perl, C#.Net) for ECU diagnosis (CAN, Flexray protocols), crash emulation and evaluation of Airbags ECU based on requirements, Rapid prototyping with Agile methodology

Education

July-2015 to July-2022

MS & PhD (Machine Learning and Computer Vision) from IIT Madras

Thesis Title: Modules for Improved Deep Learning-based Matching in Vision Tasks
Research advisor: Prof.Dr.Anurag Mittal

August-2006 to April-2010

B.E (Computer Science and Engineering) from Coimbatore Institute of Technology

Thesis: Trickle algorithm for Wireless Sensor Networks