Leveraging Photometric Consistency over Time for Sparsely Supervised Hand-Object Reconstruction

04/28/2020
by   Yana Hasson, et al.
7

Modeling hand-object manipulations is essential for understanding how humans interact with their environment. While of practical importance, estimating the pose of hands and objects during interactions is challenging due to the large mutual occlusions that occur during manipulation. Recent efforts have been directed towards fully-supervised methods that require large amounts of labeled training samples. Collecting 3D ground-truth data for hand-object interactions, however, is costly, tedious, and error-prone. To overcome this challenge we present a method to leverage photometric consistency across time when annotations are only available for a sparse subset of frames in a video. Our model is trained end-to-end on color images to jointly reconstruct hands and objects in 3D by inferring their poses. Given our estimated reconstructions, we differentiably render the optical flow between pairs of adjacent images and use it within the network to warp one frame to another. We then apply a self-supervised photometric loss that relies on the visual consistency between nearby images. We achieve state-of-the-art results on 3D hand-object reconstruction benchmarks and demonstrate that our approach allows us to improve the pose estimation accuracy by leveraging information from neighboring frames in low-data regimes.

READ FULL TEXT

page 4

page 7

page 8

page 9

research
06/06/2019

Learning Temporal Pose Estimation from Sparsely-Labeled Videos

Modern approaches for multi-person pose estimation in video require larg...
research
07/06/2020

Generative Model-Based Loss to the Rescue: A Method to Overcome Annotation Errors for Depth-Based Hand Pose Estimation

We propose to use a model-based generative loss for training hand pose e...
research
04/10/2019

H+O: Unified Egocentric Recognition of 3D Hand-Object Poses and Interactions

We present a unified framework for understanding 3D hand and object inte...
research
12/17/2020

Reconstructing Hand-Object Interactions in the Wild

In this work we explore reconstructing hand-object interactions in the w...
research
03/24/2023

DistractFlow: Improving Optical Flow Estimation via Realistic Distractions and Pseudo-Labeling

We propose a novel data augmentation approach, DistractFlow, for trainin...
research
09/01/2022

TempCLR: Reconstructing Hands via Time-Coherent Contrastive Learning

We introduce TempCLR, a new time-coherent contrastive learning approach ...
research
06/17/2021

THUNDR: Transformer-based 3D HUmaN Reconstruction with Markers

We present THUNDR, a transformer-based deep neural network methodology t...

Please sign up or login with your details

Forgot password? Click here to reset