Disentangling Semantic-to-visual Confusion for Zero-shot Learning
This work addresses the challenge of classifying unseen classes in ZSL, which is a domain-specific problem in computer vision, and is incremental as it builds on existing generative approaches with novel loss and model enhancements.
The paper tackles the problem of zero-shot learning (ZSL) image classification by addressing the semantic-to-visual confusion in generative models, proposing a multi-modal triplet loss and a disentangled class representation GAN to improve feature synthesis, resulting in superior performance on four benchmark datasets compared to state-of-the-art methods.
Using generative models to synthesize visual features from semantic distribution is one of the most popular solutions to ZSL image classification in recent years. The triplet loss (TL) is popularly used to generate realistic visual distributions from semantics by automatically searching discriminative representations. However, the traditional TL cannot search reliable unseen disentangled representations due to the unavailability of unseen classes in ZSL. To alleviate this drawback, we propose in this work a multi-modal triplet loss (MMTL) which utilizes multimodal information to search a disentangled representation space. As such, all classes can interplay which can benefit learning disentangled class representations in the searched space. Furthermore, we develop a novel model called Disentangling Class Representation Generative Adversarial Network (DCR-GAN) focusing on exploiting the disentangled representations in training, feature synthesis, and final recognition stages. Benefiting from the disentangled representations, DCR-GAN could fit a more realistic distribution over both seen and unseen features. Extensive experiments show that our proposed model can lead to superior performance to the state-of-the-arts on four benchmark datasets. Our code is available at https://github.com/FouriYe/DCRGAN-TMM.