Priyanka Patel

More about me!

Hello, my name is Priyanka Patel. I am working as an MLOps Engineer at Meshcapade since August 2024. Before joining Meshcapade, I worked as a Research Engineer at the Max Planck Institute for Intelligent Systems in Tübingen. As part of my responsibilities, I have successfully implemented scalable code for registering different body models to 3D scans. I have also trained models for 3D human pose and shape estimation from monocular images using synthetic data and developed evaluation benchmarks for 3D human pose and shape estimation methods.

Prior to my work at the Max Planck Institute, I was a Data Scientist at Zapr Media Labs in Bangalore, where I focused on anomaly detection in time series TV viewership data and user profile generation using topic modeling techniques. Before that, I have served as a Lead Engineer at the Samsung R&D Institute in Bangalore. I did my Matser's in Computer Science and Engineering from IIT Bombay and my thesis was on designing a Data-Assisted Interface for Hand-Drawn 2D Animation.

priyankapatel1201@gmail.com

My Technical Skills

Python

C++

Pytorch

OpenCV

OpenGL

Numpy

Pandas

Trimesh

Pyrender

Git

Bash

My Research Interests

Computer Vision

Computer Graphics

Deep Learning

Virtual Avatar

3D Registration

Synthetic Data

3D Human Pose and Shape Estimation

Projects and Publications

CameraHMR: Aligning People with Perspective (3DV 2025)

Priyanka Patel, Michael J. Black

We improve 3D human pose and shape estimation from monocular images by enhancing the quality of pseudo ground truth (pGT) training data. Our contributions include a novel field-of-view prediction model (HumanFoV) for estimating camera intrinsics and a dense surface keypoint detector trained on the BEDLAM dataset for more realistic body shape fitting. These advancements enable a more accurate SMPLify process and power CameraHMR, a new model achieving state-of-the-art performance.

Project Page Code arXiv Video Results

BEDLAM: A Synthetic Dataset of Bodies Exhibiting Detailed Lifelike Animated Motion (CVPR 2023)

Michael J. Black, Priyanka Patel, Joachim Tesch, Jinlong Yang

BEDLAM is a synthetic dataset consisting of realistic monocular RGB videos with accurately simulated 3D bodies, diverse clothing, and varied environments, enabling state-of-the-art accuracy in 3D human pose and shape estimation tasks using synthetic training data. As part of the project, my work involves training various regressors to estimate human pose and shape using the BEDLAM dataset. Additionally, I conducted comprehensive ablations and evaluations on different benchmarks.

Project Page Code arXiv Video

AGORA: Avatars in Geography Optimized for Regression Analysis (CVPR 2021)

Priyanka Patel, Chun-Hao P. Huang, Joachim Tesch, David T. Hoffmann, Shashank Tripathi and Michael J. Black

AGORA is a synthetic dataset that addresses the domain gap in 3D human pose and shape estimation dataset by providing highly realistic images with accurate ground truth in SMPL/SMPL-X format, including diverse poses, natural clothing, and variations in lighting and environments. It exposes the limitations of current methods and enables the development of improved models.

Project Page Code arXiv Video Evaluation

COMA in Pytorch

COMA (Convolutional Mesh Autoencoder) is an advanced technique for accurately representing 3D faces in a low-dimensional latent space. It surpasses the performance of linear PCA models while employing fewer parameters. In order to make it more accessible and adaptable, I have successfully ported the original code from Tensorflow to PyTorch, leveraging the capabilities of the torch-geometric library.

Project Page Code

Model Registration Pipeline for 3D Scans

A modular and scalable pipeline for registering different body models, such as SMPL, SMPL-X, and SMIL, to 3D scans in Pytorch. The key features include rendering 3D scans from multiple camera viewpoints, estimating 2D keypoints in each view, optimizing model poses by minimizing the projected 2D keypoints loss across all views, and utilizing point-to-surface distance for precise alignment of the model and the scan after an initial proximity estimation.

Registration for clothed 3D Scans

As part of the project AGORA, I work on registering SMPL-X model to 3D scans with clothing and hair. This task presents a greater challenge compared to scans with minimal clothing, as the point-to-surface distance tends to distort the body in order to accommodate the clothing. To address this issue, I employ Graphonomy, which help in separating the scan into skin and clothing vertices. I further implemented an optimization term to minimize the point to surface distance for skin region while simultaneously ensuring the body remains appropriately enclosed within the clothing.