Multiple instance learning on deep features for weakly supervised object detection with extreme domain shifts

TL;DR

This paper demonstrates that a straightforward multiple instance learning approach on pre-trained deep features can effectively perform weakly supervised object detection across diverse and challenging non-photographic datasets without fine-tuning.

Contribution

It introduces a simple, fine-tuning-free multiple instance learning method that generalizes well to various non-photographic domains and unseen classes.

Findings

Achieves competitive results on paintings, watercolors, cliparts, and comics datasets.

No fine-tuning or cross-domain learning required, enabling quick adaptation.

Effective for new visual categories without extensive training.

Abstract

Weakly supervised object detection (WSOD) using only image-level annotations has attracted a growing attention over the past few years. Whereas such task is typically addressed with a domain-specific solution focused on natural images, we show that a simple multiple instance approach applied on pre-trained deep features yields excellent performances on non-photographic datasets, possibly including new classes. The approach does not include any fine-tuning or cross-domain learning and is therefore efficient and possibly applicable to arbitrary datasets and classes. We investigate several flavors of the proposed approach, some including multi-layers perceptron and polyhedral classifiers. Despite its simplicity, our method shows competitive results on a range of publicly available datasets, including paintings (People-Art, IconArt), watercolors, cliparts and comics and allows to quickly learn unseen visual categories.