Zero-Shot Object-Centric Representation Learning
This work addresses the limitation of object-centric methods being applied only in-distribution, enabling more general-purpose models for visual scene decomposition, though it is incremental as it builds on existing pre-trained features and fine-tuning techniques.
The paper tackled the problem of object-centric representation learning lacking zero-shot generalization by introducing a benchmark of eight datasets and analyzing factors affecting transferability. They proposed a fine-tuning strategy for pre-trained vision encoders, achieving state-of-the-art performance in unsupervised object discovery with strong zero-shot transfer to unseen datasets.
The goal of object-centric representation learning is to decompose visual scenes into a structured representation that isolates the entities. Recent successes have shown that object-centric representation learning can be scaled to real-world scenes by utilizing pre-trained self-supervised features. However, so far, object-centric methods have mostly been applied in-distribution, with models trained and evaluated on the same dataset. This is in contrast to the wider trend in machine learning towards general-purpose models directly applicable to unseen data and tasks. Thus, in this work, we study current object-centric methods through the lens of zero-shot generalization by introducing a benchmark comprising eight different synthetic and real-world datasets. We analyze the factors influencing zero-shot performance and find that training on diverse real-world images improves transferability to unseen scenarios. Furthermore, inspired by the success of task-specific fine-tuning in foundation models, we introduce a novel fine-tuning strategy to adapt pre-trained vision encoders for the task of object discovery. We find that the proposed approach results in state-of-the-art performance for unsupervised object discovery, exhibiting strong zero-shot transfer to unseen datasets.