Timezone: »

 
Poster
Compositional Zero-Shot Learning via Fine-Grained Dense Feature Composition
Dat Huynh · Ehsan Elhamifar

Wed Dec 09 09:00 AM -- 11:00 AM (PST) @ Poster Session 3 #1006

We develop a novel generative model for zero-shot learning to recognize fine-grained unseen classes without training samples. Our observation is that generating holistic features of unseen classes fails to capture every attribute needed to distinguish small differences among classes. We propose a feature composition framework that learns to extract attribute-based features from training samples and combines them to construct fine-grained features for unseen classes. Feature composition allows us to not only selectively compose features of unseen classes from only relevant training samples, but also obtain diversity among composed features via changing samples used for composition. In addition, instead of building a global feature of an unseen class, we use all attribute-based features to form a dense representation consisting of fine-grained attribute details. To recognize unseen classes, we propose a novel training scheme that uses a discriminative model to construct features that are subsequently used to train itself. Therefore, we directly train the discriminative model on composed features without learning separate generative models. We conduct experiments on four popular datasets of DeepFashion, AWA2, CUB, and SUN, showing that our method significantly improves the state of the art.

Author Information

Dat Huynh (Northeastern University)

My main interest lies in reducing the amount of annotation needed to train deep learning systems for visual recognition tasks such as object detection and segmentation. Specifically, I am mesmerized by how much machines can learn with a few or zero training sample(s), and with missing annotations in weakly-supervised setting. If you are interested in my research or collaboration, I can be reached via: Email: huynh [dot] dat [at] northeastern [dot] edu

Ehsan Elhamifar (Northeastern University)

More from the Same Authors