English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recently, Mix-style data augmentation methods (e.g., Mixup and CutMix) haveshown promising performance in various visual tasks. However, these methods areprimarily designed for single-label images, ignoring the considerablediscrepancies between single- and multi-label images, i.e., a multi-label imageinvolves multiple co-occurred categories and fickle object scales. On the otherhand, previous multi-label image classification (MLIC) methods tend to designelaborate models, bringing expensive computation. In this paper, we introduce asimple but effective augmentation strategy for multi-label imageclassification, namely SpliceMix. The "splice" in our method is two-fold: 1)Each mixed image is a splice of several downsampled images in the form of agrid, where the semantics of images attending to mixing are blended withoutobject deficiencies for alleviating co-occurred bias; 2) We splice mixed imagesand the original mini-batch to form a new SpliceMixed mini-batch, which allowsan image with different scales to contribute to training together. Furthermore,such splice in our SpliceMixed mini-batch enables interactions between mixedimages and original regular images. We also offer a simple and non-parametricextension based on consistency learning (SpliceMix-CL) to show the flexibleextensibility of our SpliceMix. Extensive experiments on various tasksdemonstrate that only using SpliceMix with a baseline model (e.g., ResNet)achieves better performance than state-of-the-art methods. Moreover, thegeneralizability of our SpliceMix is further validated by the improvements incurrent MLIC methods when married with our SpliceMix. The code is available athttps://github.com/zuiran/SpliceMix.

SpliceMix: A Cross-scale and Semantic Blending Augmentation Strategy for Multi-label Image Classification