Attention-Guided Masked Autoencoders For Learning Image Representations
This work addresses the need for more robust and object-centric pre-training in computer vision, though it is incremental as it builds on existing MAE methods.
The paper tackles the problem of learning object-focused image representations in masked autoencoders by introducing an attention-guided loss function, resulting in improved linear probing and k-NN classification results on benchmarks and increased robustness against varying backgrounds.
Masked autoencoders (MAEs) have established themselves as a powerful method for unsupervised pre-training for computer vision tasks. While vanilla MAEs put equal emphasis on reconstructing the individual parts of the image, we propose to inform the reconstruction process through an attention-guided loss function. By leveraging advances in unsupervised object discovery, we obtain an attention map of the scene which we employ in the loss function to put increased emphasis on reconstructing relevant objects, thus effectively incentivizing the model to learn more object-focused representations without compromising the established masking strategy. Our evaluations show that our pre-trained models learn better latent representations than the vanilla MAE, demonstrated by improved linear probing and k-NN classification results on several benchmarks while at the same time making ViTs more robust against varying backgrounds.