How PARTs assemble into wholes: Learning the relative composition of images
This addresses the problem of improving spatial understanding in self-supervised learning for researchers and practitioners, offering a novel approach with broad applications, though it builds incrementally on existing pretext tasks.
The paper tackles the limitation of grid-based self-supervised learning methods in capturing continuous object compositions by introducing PART, which uses relative transformations between off-grid patches, resulting in outperforming methods like MAE and DropPos in tasks like object detection and time series prediction.
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 process that generalizes beyond occlusions and deformations. In tasks requiring precise spatial understanding such as object detection and time series prediction, PART outperforms strong grid-based methods like MAE and DropPos, while also maintaining competitive performance on global classification tasks with minimal hyperparameter tuning. By breaking free from grid constraints, PART opens up an exciting new trajectory for universal self-supervised pretraining across diverse datatypes-from natural images to EEG signals-with promising potential in video, medical imaging, and audio.