Sketch-a-Classifier: Sketch-based Photo Classifier Generation

TL;DR

This paper introduces Sketch-a-Classifier, a method that synthesizes photo classifiers directly from free-hand sketches, enabling rapid, user-driven image recognition model creation without requiring annotated photos or category descriptions.

Contribution

It presents a category-agnostic model regression approach that maps sketches to classifiers, allowing new categories to be recognized without training photos or attribute descriptions.

Findings

Effective classifier synthesis from sketches demonstrated

Enhances classifier granularity and complements zero-shot learning

No need for annotated photos for new categories

Abstract

Contemporary deep learning techniques have made image recognition a reasonably reliable technology. However training effective photo classifiers typically takes numerous examples which limits image recognition's scalability and applicability to scenarios where images may not be available. This has motivated investigation into zero-shot learning, which addresses the issue via knowledge transfer from other modalities such as text. In this paper we investigate an alternative approach of synthesizing image classifiers: almost directly from a user's imagination, via free-hand sketch. This approach doesn't require the category to be nameable or describable via attributes as per zero-shot learning. We achieve this via training a {model regression} network to map from {free-hand sketch} space to the space of photo classifiers. It turns out that this mapping can be learned in a category-agnostic way, allowing photo classifiers for new categories to be synthesized by user with no need for annotated training photos. {We also demonstrate that this modality of classifier generation can also be used to enhance the granularity of an existing photo classifier, or as a complement to name-based zero-shot learning.