Joint-Embedding Predictive Architecture for Self-Supervised Learning of Mask Classification Architecture

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Dong-Hee, Cho, Sungduk, Cho, Hyeonwoo, Park, Chanmin, Kim, Jinyoung, Kim, Won Hwa
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917722261028864
author Kim, Dong-Hee
Cho, Sungduk
Cho, Hyeonwoo
Park, Chanmin
Kim, Jinyoung
Kim, Won Hwa
author_facet Kim, Dong-Hee
Cho, Sungduk
Cho, Hyeonwoo
Park, Chanmin
Kim, Jinyoung
Kim, Won Hwa
contents In this work, we introduce Mask-JEPA, a self-supervised learning framework tailored for mask classification architectures (MCA), to overcome the traditional constraints associated with training segmentation models. Mask-JEPA combines a Joint Embedding Predictive Architecture with MCA to adeptly capture intricate semantics and precise object boundaries. Our approach addresses two critical challenges in self-supervised learning: 1) extracting comprehensive representations for universal image segmentation from a pixel decoder, and 2) effectively training the transformer decoder. The use of the transformer decoder as a predictor within the JEPA framework allows proficient training in universal image segmentation tasks. Through rigorous evaluations on datasets such as ADE20K, Cityscapes and COCO, Mask-JEPA demonstrates not only competitive results but also exceptional adaptability and robustness across various training scenarios. The architecture-agnostic nature of Mask-JEPA further underscores its versatility, allowing seamless adaptation to various mask classification family.
format Preprint
id arxiv_https___arxiv_org_abs_2407_10733
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Joint-Embedding Predictive Architecture for Self-Supervised Learning of Mask Classification Architecture
Kim, Dong-Hee
Cho, Sungduk
Cho, Hyeonwoo
Park, Chanmin
Kim, Jinyoung
Kim, Won Hwa
Computer Vision and Pattern Recognition
In this work, we introduce Mask-JEPA, a self-supervised learning framework tailored for mask classification architectures (MCA), to overcome the traditional constraints associated with training segmentation models. Mask-JEPA combines a Joint Embedding Predictive Architecture with MCA to adeptly capture intricate semantics and precise object boundaries. Our approach addresses two critical challenges in self-supervised learning: 1) extracting comprehensive representations for universal image segmentation from a pixel decoder, and 2) effectively training the transformer decoder. The use of the transformer decoder as a predictor within the JEPA framework allows proficient training in universal image segmentation tasks. Through rigorous evaluations on datasets such as ADE20K, Cityscapes and COCO, Mask-JEPA demonstrates not only competitive results but also exceptional adaptability and robustness across various training scenarios. The architecture-agnostic nature of Mask-JEPA further underscores its versatility, allowing seamless adaptation to various mask classification family.
title Joint-Embedding Predictive Architecture for Self-Supervised Learning of Mask Classification Architecture
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.10733