Shankar Chatterjee **** */* Clinton Street

Los Angeles, CA 90004

323-***-**** (mobile)

ad19fu@r.postjobfree.com

Areas of Expertise

Deep Learning, Generative AI, Contextual Intelligence, Computer Vision, Scale-agnostic Semantic Image/Video Recognition, Object detection, Image quality, Segmentation, Scene Reconstruction.

Creative applications of the above in film production, on-line digital ad insertion, surveillance, augmented reality(AR) etc.

Professional Experience

Computer Vision Consultant July 2019 -

• Development of car license plate recognition, exterior body work inspection, etc., detection of wheelchairs, strollers etc., on a cpu box.

• Facial expression swapping on videos using few shots generative adversial network(FSGAN). Senior Perception Engineer Feb 2021 - May 2023

Miso Robotics, Inc., Pasadena CA

• Detection of steaks(cook status) on hot-plate using deep learning.

• Calibration of cameras (Intel D35) using high resolution OptiTrack II camera system. Senior Computer Vision Scientist Oct 2015 - June 2019 Gumgum Inc., Santa Monica CA

• Research/Development of deep-learning based image/video recognition, object detection (training, model fine-tuning), clarity estimation, face detection, facial model(3Dmm) generation, instance seg- mentation, tracking, brand safety, etc.

• Object detectors explored: SSD, DSOD, YOLOv3/4 etc. Segmentation networks: PSPNet, CCNet, FC8, DeepLabV3+ etc. Classifiers : AlexNet, ResNet50, SENet, EfficientNet etc.

• Application of Convolutional Neural Network(CNN) in human body/face pose estimation, facial emo- tion recognition, hair/skin/clothes segmentation etc.

• Conceptualization, pipeline development and implementation of automated virtual ad placement in images/videos.

• Frame-accurate and 3D-consistent dynamic ad insertion (DAI) in (sports) video using contextual intel- ligence (for OTT).

• Frameworks used: Caffe(/2), MxNet, PyTorch, Keras/TensorFlow.

• Deployment of products to AWS EC2 cloud for on-line ad-serving. Senior Computer Vision Engineer Nov 2014 - June 2015 FaceCake Marketing Technologies, Calabasas, CA

• Development of original algorithms for digital fashion products, face recognition and pose estimation. Research Consultant Nov 2012 - Sept 2014

• Development of original algorithms for body skeleton, face keypoints extraction.

• Development of seamless image/video stitching software for VR applications. Senior Stereo Algorithm Engineer Jan 2012 - Sept 2012 Digital Domain Media Group, Venice, CA

• Development of original algorithms for stereo conversion, digital inpaintiong.

• Development of geometric solutions for 3D scene recosntructions(static). Senior Algorithm Developer Aug 2009 - Nov 2011

PJF Productions, Studio City, CA

• Reimplementation of old Pacific Title software codebase, new optical flow (variational) algorithms for grain management.

Senior Algorithmist July 2008 - June 2009

Reliance MediaWorks, Burbank, CA

• Design and implementation of cel-dirt detection/removal algorithm, graphcut-based image segmenta- tion, optical flow for motion estimation.

Graphics Scientist (consultant) April 2008 - June 2008 Rhythm & Hues Studio, Los Angeles, CA

• Design and implementation of a nonlinear multigrid methods for motion esimation. Principal Research Scientist May 2003 - March 2008 Pacific Title & Arts Studio, Hollywood, CA

• Design of software pipeline for Restoration/Digital Intermediate workflow. Also, principal contributor to the development(in-house) with innovative algorithms and efficient computational architecture for automated dust-busting, scratch detection/repair, film grain management, spatial/temporal deflicker, color fading and other artifacts.

Senior Research Scientist April 1996 - May 2003

Cinesite - A Kodak Company, Hollywood, CA

• Research, algorithm design and original implementations of various film processing tasks, e.g., optical flow and its applications, feature tracking (both 2D and 3D), camera calibration (internal, external) and tracking, image warping/morphing, view-morphing(3D), depth-of-field and motion blur, frame stabilization, wire-removal, object editing, etc.

• Screen Credits (sample) The Lost Souls, Thirteen Days, Hired Hand (restoration), Rear Window (restoration). Research Consultant June 1995 - Jan 1996

Autonetics Division, Rockwell International, Anaheim, CA

• Design of vision-based advanced occupant sensors (AOSM),e.g, car-seats, humans.

• Implementation of vehicle counting algorithm from video. Research Scientist July 1994 - May 1996

Visual Computing Lab., Univ. of California, San Diego, La Jolla, CA

• Research on similarity measures for image database (e.g., face, textures) indexing and content-based retrieval, image coding using wavelets, VQ and space-filling curves. Assistant Professor Aug. 1987 - June 1994.

Dept. of Elec. & Comp. Engg., Univ. of California, San Diego, La Jolla, CA

• Involved in building an image-processing research laboratory with Elec. & Comp. Engg. Deptt., UCSD.

• Initiated an image-analysis laboratory with Dept. of Neurosciences, UCSD for 3D-visualization from scanned serial sections of hippocampus.

• Collaborated in research with UCSD Dept. of Opthalmology in automated structural analysis (STARE) and indexing of retinal image database. Developed appropriate tools for disease/manifestation paradigm. Education

Ph.D. in Electrical Engineering-Systems

University of Southern California, Los Angeles, CA Thesis Title: Synthesis, Analysis and Segmentation of Image Textures M.S. in Electrical Engineering

Louisiana State University, Baton Rouge, LA

B.Tech. (Hons.) in Electronics & Elect. Comm. Engg. Indian Institute of Technology, Kharagpur, India

Related Skills

• Software stack: C/C++, Python.

• Familiar with : Linux (ubuntu), Mac OSX.

Caffe, MxNet, PyTorch. TensorFlow

Github/Gitflow, Docker, Kubernetes.

OpenCV, PCL, OpenGL, Blender 3D, numpy, scipy, scikit-learn. Jira, Confluence, Slack, AWS(EC2, S3, Media Tailor), Drone CI/CD. Several compositing packages - Cineon, Nuke, Polaris (in-house) etc. Awards/Honors

• Best poster paper, Asian Machine Vision (AMV2018) workshop, Perth (for work on Mark- erless Augmented Advertising in Sports Videos), 2018.

• Best Applications paper, System Society of India, 1991 (for work on robust motion esti- mation).

• JSEP Research Assistant, Univ. of So. Calif., Los Angeles, CA, 1982-1986.

• Reseacrh Assistant, Dept. of Elec. Engg., Louisiana State Univ., Baton Rouge, La., 1979- 1981.

• National Merit Scholarship, Govt. of India, 1973-1978. Sample Publications

• J. Mammen, S. Chaudhuri and S. Chatterjee, “Line Scratch Removal from Digitized Mo- tion Picture Frames”, Proc. Intl. Conf. Multimedia Processing and Systems (ICMPS), IIT Madras, Aug. 2000.

• S. Chaudhuri, S. Sharma and S. Chatterjee, “Robust and Recursive Estimation of Motion Parameters”, Computer Vision, Graphics and Image Processing - Image Understanding, vol. 64, no. 3, Nov. 1996.

• C.C. Chang and S. Chatterjee, “Some Deterministic Approximations for Stochastic Stereo Matching”, Signal Processing, vol.45, no.2, pp. 143-160, Aug. 1995.

• S. Chatterjee, “Texture Synthesis and Classification”, in Markov Random Fields: Theory and Applications, eds. R. Chellappa and A.K. Jain, Academic Press, pp. 159-177, 1992.

• S. Chaudhuri and S. Chatterjee, “Performance Analysis of Linear, Least-Squares Methods in 3-D Motion Estimation”, IEEE Trans. on Robotics & Automation, vol. 7, no. 8, pp. 707- 714, Oct., 1991.

Contact this candidate