Projects

Project List Sony 3DCG R&D : OnSpecta : 2017-present : 2012-2017 : 2007-2012 : 2004-2007 : 2000-2004

Sony
  • 3D Volumetric Performance Capture link
Onspecta
  • BERT Optimization
    • Pytorch Vs TensorFlow link
    • Benchmarking link
    • Architecture link
  • Weight Pruning link
  • Resnet50 optimization link
  • Quantization link
Washington State University
  • NSVF link
  • Gan based controllable pose and dress link
  • Segment videos based on events and classify them link
  • Jpg image classifer
  • Compression experiments link
  • Combinatorial experiments link
  • ML tutorials - list of experiments to understand various types of networks and different ml paradigms link
  • Computational genomics link
  • AGI - brain project
  • Video compression - GAN based, progressive GAN, VQ-VAE-2
  • In Media Product Placement - DensePose, 3d-kenburns, detectron2, InGAN, SinGAN, StyleGAN-2, vid2vid, few-shot-vid2vid
  • multi agent RL
  • Pose Comparison
  • Deep Neural Network Inference Optimization:
    • Transformer Quantization - Classification and Seq
    • Pytorch Quantization - Tensorflow Quantization
    • Pruning
    • ARM platform
    • ResNet Quantization
  • Virtual Studio - ADGAN, CPISADGAN, Image HArmonization, LIP_SSL, Human Segmentation, Single-Human-Parsing-LIP,
  • Nerfies, PoseTransfer
  • vdo mocap -
  • HumanModels - PifuHD, CAPE, OpenPose, HairNet,
  • Multi view rendering
    • Cloth hair pose gen - ADGAN, PArsing-R-CNN, pi-GAN-pytorch, tex2shape
    • colmap, DeepSDF, nerf,
    • nerf_universe - kilonerf, NSVF, pixel-nerf, nerf in the wild, [code] - Pose based nerf - densepose-siren, image_interp, pose2d_cnn, pose2d_gen, pose2d_siren, syn_data_train
    • NeRF++, NSVF, NSVF_update
    • neural_sparse_voxel_field - 1D signal encoding, analysis, fourier_feature_experiment, image_encoding, mesh_generation
  • Vdo_mocap - dancebabydance
    • detectron2, Detr, EasyMocap, FRankMocap, human_body_prior,
    • nerf++, neuralbody, pifuhd, pixel2style2pixel
    • pose_to_mocap, video2bvh, VIBE, SPIN, SMPL, SMPLX, SMPLify-X, SMPLicit
    • theta_beta_iteration
    • MeshTransformer
  • video_pose_estimation
  • videoavatars, VideoPose3D, White-Box-Cartoonization - Cartoon Generation, Dress extractor, Vtuber
Polycom
  • Video Conferencing plugin} for chrome 64 bit using PPAPI
  • RTP, RDP and HTML5 based Content Collaboration} between MCU, Lync and Browser endpoints
  • Automation Platform for bringing up Infrastructure for Video as a service on VMWare
  • Platform Director: Life cycle management of virtual instances on VMWare using viJava
Ittiam Systems
  • Mpeg2 (MP@HL) HD(1080i@30fps) Encoder/Transcoder for Broadcast on Netra (IVAHD Accel)
  • MPEG4 Simple Profile (SP) HD(720p@30fps) Encoder for Smart Phones on OMAP34xx (Arm+DSP+IVA)
  • Mpeg4 SP D1 (480p@30fps) Encoder for Portable Media Player and Recorder on OMAP3430.
  • Mpeg4 Advanced Simple Profile HD Decoder on OMAP3430
  • Mpeg4, H.263 and Mpeg2 Encoders for catalogue products on C64x+ DSP
  • Handling 3:2 pulldown by reducing frame rate and generation of Timestamp information
  • Migrating the mpeg4 ME optimizations to MPEG2 non-accelerated workspace
  • Adding MV cost for Motion Estimation [MPEG4 non-accelerated]
  • H263 Encoder and its extensions D1 on C64x+ DSP
  • Added XDM 1.0 API support for Mpeg4 non accelerated version
  • Optimization of MPEG4 non accelerated encoder
Tata Elxsi
  • ‘N’ mb level optimization of core loop Difference+DCT+Q+IQ+IDCT.
  • Study of SD H.264 Scalable Video Codec on SCP
  • Real Time HD (High Definition) MPEG2 Video Encoder on Multiple SCP (S5) Processors
  • Real Time HD H.264 Video Encoder On Multiple SCP (S5) Processors
  • H.264 Encoder modules development
  • Real Time Preview of HD bit streams on PC
  • Study on video transcoding of HD MPEG2 bit streams to SD (Standard Definition) bit streams
Anna University
  • Visualization of White Matter Fiber Tract from DT-MR (Diffusion Tensor Magnetic Resonance) Image.
  • Micro-Controller based Personal Data Assistant using MC89C51
Kaggle projects