SketchEmbedNet: Learning Novel Concepts by Imitating Drawings
This work addresses the challenge of learning robust visual representations for novel concepts, though it appears incremental as it builds on prior sketch generation methods.
The paper tackled the problem of learning image representations by training a model to generate sketches, showing that this approach produces informative embeddings for novel examples, classes, and datasets in a few-shot setting, with demonstrated structure and compositionality.
Sketch drawings capture the salient information of visual concepts. Previous work has shown that neural networks are capable of producing sketches of natural objects drawn from a small number of classes. While earlier approaches focus on generation quality or retrieval, we explore properties of image representations learned by training a model to produce sketches of images. We show that this generative, class-agnostic model produces informative embeddings of images from novel examples, classes, and even novel datasets in a few-shot setting. Additionally, we find that these learned representations exhibit interesting structure and compositionality.