How PARTs assemble into wholes: Learning the relative composition of images

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ayoughi, Melika, Abnar, Samira, Huang, Chen, Sandino, Chris, Lala, Sayeri, Dhekane, Eeshan Gunesh, Busbridge, Dan, Zhai, Shuangfei, Thilak, Vimal, Susskind, Josh, Mettes, Pascal, Groth, Paul, Goh, Hanlin
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915673693749248
author Ayoughi, Melika
Abnar, Samira
Huang, Chen
Sandino, Chris
Lala, Sayeri
Dhekane, Eeshan Gunesh
Busbridge, Dan
Zhai, Shuangfei
Thilak, Vimal
Susskind, Josh
Mettes, Pascal
Groth, Paul
Goh, Hanlin
author_facet Ayoughi, Melika
Abnar, Samira
Huang, Chen
Sandino, Chris
Lala, Sayeri
Dhekane, Eeshan Gunesh
Busbridge, Dan
Zhai, Shuangfei
Thilak, Vimal
Susskind, Josh
Mettes, Pascal
Groth, Paul
Goh, Hanlin
contents The composition of objects and their parts, along with object-object positional relationships, provides a rich source of information for representation learning. Hence, spatial-aware pretext tasks have been actively explored in self-supervised learning. Existing works commonly start from a grid structure, where the goal of the pretext task involves predicting the absolute position index of patches within a fixed grid. However, grid-based approaches fall short of capturing the fluid and continuous nature of real-world object compositions. We introduce PART, a self-supervised learning approach that leverages continuous relative transformations between off-grid patches to overcome these limitations. By modeling how parts relate to each other in a continuous space, PART learns the relative composition of images-an off-grid structural relative positioning that is less tied to absolute appearance and can remain coherent under variations such as partial visibility or stylistic changes. In tasks requiring precise spatial understanding such as object detection and time series prediction, PART outperforms grid-based methods like MAE and DropPos, while maintaining competitive performance on global classification tasks. By breaking free from grid constraints, PART opens up a new trajectory for universal self-supervised pretraining across diverse datatypes-from images to EEG signals-with potential in medical imaging, video, and audio.
format Preprint
id arxiv_https___arxiv_org_abs_2506_03682
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle How PARTs assemble into wholes: Learning the relative composition of images
Ayoughi, Melika
Abnar, Samira
Huang, Chen
Sandino, Chris
Lala, Sayeri
Dhekane, Eeshan Gunesh
Busbridge, Dan
Zhai, Shuangfei
Thilak, Vimal
Susskind, Josh
Mettes, Pascal
Groth, Paul
Goh, Hanlin
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
The composition of objects and their parts, along with object-object positional relationships, provides a rich source of information for representation learning. Hence, spatial-aware pretext tasks have been actively explored in self-supervised learning. Existing works commonly start from a grid structure, where the goal of the pretext task involves predicting the absolute position index of patches within a fixed grid. However, grid-based approaches fall short of capturing the fluid and continuous nature of real-world object compositions. We introduce PART, a self-supervised learning approach that leverages continuous relative transformations between off-grid patches to overcome these limitations. By modeling how parts relate to each other in a continuous space, PART learns the relative composition of images-an off-grid structural relative positioning that is less tied to absolute appearance and can remain coherent under variations such as partial visibility or stylistic changes. In tasks requiring precise spatial understanding such as object detection and time series prediction, PART outperforms grid-based methods like MAE and DropPos, while maintaining competitive performance on global classification tasks. By breaking free from grid constraints, PART opens up a new trajectory for universal self-supervised pretraining across diverse datatypes-from images to EEG signals-with potential in medical imaging, video, and audio.
title How PARTs assemble into wholes: Learning the relative composition of images
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2506.03682