Do MLLMs Exhibit Human-like Perceptual Behaviors? HVSBench: A Benchmark for MLLM Alignment with Human Perceptual Behavior

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Jiaying, Ye, Shuquan, Xu, Dan, Ouyang, Wanli, Lau, Rynson W. H.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912769819803648
author Lin, Jiaying
Ye, Shuquan
Xu, Dan
Ouyang, Wanli
Lau, Rynson W. H.
author_facet Lin, Jiaying
Ye, Shuquan
Xu, Dan
Ouyang, Wanli
Lau, Rynson W. H.
contents While Multimodal Large Language Models (MLLMs) excel at many vision tasks, it is unknown if they exhibit human-like perceptual behaviors. To evaluate this, we introduce HVSBench, the first large-scale benchmark with over 85,000 samples designed to test MLLM alignment with the human visual system (HVS). The benchmark covers 13 categories across 5 key fields: Prominence, Subitizing, Prioritizing, Free-Viewing, and Searching. Our comprehensive evaluation reveals a significant perceptual gap: even state-of-the-art MLLMs achieve only moderate results. In contrast, human participants demonstrate strong performance, significantly outperforming all models. This underscores the high quality of HVSBench and the need for more human-aligned AI. We believe our benchmark will be a critical tool for developing the next generation of explainable MLLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2412_09603
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Do MLLMs Exhibit Human-like Perceptual Behaviors? HVSBench: A Benchmark for MLLM Alignment with Human Perceptual Behavior
Lin, Jiaying
Ye, Shuquan
Xu, Dan
Ouyang, Wanli
Lau, Rynson W. H.
Computer Vision and Pattern Recognition
While Multimodal Large Language Models (MLLMs) excel at many vision tasks, it is unknown if they exhibit human-like perceptual behaviors. To evaluate this, we introduce HVSBench, the first large-scale benchmark with over 85,000 samples designed to test MLLM alignment with the human visual system (HVS). The benchmark covers 13 categories across 5 key fields: Prominence, Subitizing, Prioritizing, Free-Viewing, and Searching. Our comprehensive evaluation reveals a significant perceptual gap: even state-of-the-art MLLMs achieve only moderate results. In contrast, human participants demonstrate strong performance, significantly outperforming all models. This underscores the high quality of HVSBench and the need for more human-aligned AI. We believe our benchmark will be a critical tool for developing the next generation of explainable MLLMs.
title Do MLLMs Exhibit Human-like Perceptual Behaviors? HVSBench: A Benchmark for MLLM Alignment with Human Perceptual Behavior
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.09603