Details

Autor(en) / Beteiligte

• Dornaika, F.
• Sun, D.
• Hammoudi, K.
• Charafeddine, J.
• Cabani, A.
• Zhang, C.

Titel

Object-centric Contour-aware Data Augmentation Using Superpixels of Varying Granularity

Ist Teil von

• Pattern recognition, 2023-07, Vol.139, p.109481, Article 109481

Ort / Verlag

Elsevier Ltd

Erscheinungsjahr

2023

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• •We point out the potential drawbacks of existing region-based dropout augmentation strategies.•We present an attention-driven, contour-aware CutMix method for data augmentation.•It can capture the most discriminative contour-preserving regions and significantly outperforms existing CutMix-based augmentation methods.•We propose a tradeoff between augmentation diversification and regularization concentration by deriving the relevant superpixels with different granularities.•We conduct extensive experiments with different image datasets and different CNN backbones. Regional dropout strategies have demonstrated to be very effective in improving both the performance and the generalization capability of deep learning models. However, when such strategies are performed in a totally random manner, the background noise and label mismatch problems arise. To tackle such problems, existing approaches typically focus on regions with the highest distinctiveness. Yet, there are two main drawbacks of existing approaches: (I) Many existing region-based augmentation methods can only use rectangular regions, resulting in the loss of object contour information; (II) Deterministic selection of the most discriminative regions leads to poor diversification in data augmentation. In fact, a trade-off is needed between diversification and concentration, which can decrease the undesirable noise. In this paper, we propose a novel object-centric contour-aware CutMix data augmentation strategy with arbitrary- shape and size superpixel supports, which is hereafter referred to as OcCaMix for short. It not only captures the most discriminative regions, but also effectively preserves the contour details of the objects. Moreover, it enables the search of natural object parts of different sizes. Extensive experiments on a large number of benchmark datasets show that OcCaMix significantly outperforms state-of-the-art CutMix based data augmentation methods in classification tasks. The source codes and trained models are available at https://github.com/DanielaPlusPlus/OcCaMix.