Saved in:
Bibliographic Details
Main Authors: Lin, Ethan, Zhao, Linxi, Sehgal, Atharva, Sun, Jennifer J.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.03542
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913933125746688
author Lin, Ethan
Zhao, Linxi
Sehgal, Atharva
Sun, Jennifer J.
author_facet Lin, Ethan
Zhao, Linxi
Sehgal, Atharva
Sun, Jennifer J.
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2507_03542
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Beyond Accuracy: Metrics that Uncover What Makes a 'Good' Visual Descriptor
Lin, Ethan
Zhao, Linxi
Sehgal, Atharva
Sun, Jennifer J.
Computer Vision and Pattern Recognition
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.
title Beyond Accuracy: Metrics that Uncover What Makes a 'Good' Visual Descriptor
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.03542