Swapping Semantic Contents for Mixing Images
This addresses the bottleneck of labeled data availability in low-label settings, offering an incremental improvement over existing mixing data augmentations.
The paper tackles the problem of limited labeled data in semi-supervised learning by introducing the SciMix framework, which generates new labeled samples by swapping semantic contents in images, resulting in improved performance for methods like Mean Teacher and Fixmatch.
Deep architecture have proven capable of solving many tasks provided a sufficient amount of labeled data. In fact, the amount of available labeled data has become the principal bottleneck in low label settings such as Semi-Supervised Learning. Mixing Data Augmentations do not typically yield new labeled samples, as indiscriminately mixing contents creates between-class samples. In this work, we introduce the SciMix framework that can learn to generator to embed a semantic style code into image backgrounds, we obtain new mixing scheme for data augmentation. We then demonstrate that SciMix yields novel mixed samples that inherit many characteristics from their non-semantic parents. Afterwards, we verify those samples can be used to improve the performance semi-supervised frameworks like Mean Teacher or Fixmatch, and even fully supervised learning on a small labeled dataset.