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

Nicolas Gonthier (1, 2) , Saïd Ladjal (1, 2) , Yann Gousseau (1, 2)

1 IMAGES - Image, Modélisation, Analyse, GEométrie, Synthèse
2 IDS - Département Images, Données, Signal

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.

Mots clés: Deep learning; Convolutional neural networks; Weakly supervised object detection; Non-photographic images; Art analysis; Multiple instance learning

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]

Nicolas Gonthier, Saïd Ladjal, Yann Gousseau. Multiple instance learning on deep features for weakly supervised object detection with extreme domain shifts. Computer Vision and Image Understanding, 2022, 214, ⟨10.1016/j.cviu.2021.103299⟩. ⟨hal-03210265⟩