Beyond Accuracy: Metrics that Uncover What Makes a 'Good' Visual Descriptor

TL;DR

This paper introduces new metrics to evaluate the quality of visual descriptors, focusing on their representational capacity and alignment with model pre-training data, providing insights beyond traditional accuracy measures.

Contribution

It proposes two novel alignment-based metrics, Global Alignment and CLIP Similarity, to better understand descriptor effectiveness in vision-language models.

Findings

Metrics reveal how descriptor strategies interact with model properties

Descriptors' effectiveness depends on semantic clarity and pre-training data presence

Evaluation methods extend beyond accuracy to assess descriptor quality

Abstract

Text-based visual descriptors--ranging from simple class names to more descriptive phrases--are widely used in visual concept discovery and image classification with vision-language models (VLMs). Their effectiveness, however, depends on a complex interplay of factors, including semantic clarity, presence in the VLM's pre-training data, and how well the descriptors serve as a meaningful representation space. In this work, we systematically analyze descriptor quality along two key dimensions: (1) representational capacity, and (2) relationship with VLM pre-training data. We evaluate a spectrum of descriptor generation methods, from zero-shot LLM-generated prompts to iteratively refined descriptors. Motivated by ideas from representation alignment and language understanding, we introduce two alignment-based metrics--Global Alignment and CLIP Similarity--that move beyond accuracy. These metrics shed light on how different descriptor generation strategies interact with foundation model properties, offering new ways to study descriptor effectiveness beyond accuracy evaluations.