An official website of the United States government
Medical vision-language models (VLMs) combine computer vision (CV) and natural language processing (NLP) to analyze visual and textual medical data. Our paper reviews recent advancements in developing VLMs specialized for healthcare, focusing on publicly available models designed for medical report generation and visual question answering (VQA). We provide background on NLP and CV, explaining how techniques from both fields are integrated into VLMs, with visual and language data often fused using Transformer-based architectures to enable effective learning from multimodal data. Key areas we address include the exploration of 18 public medical vision-language datasets, in-depth analyses of the architectures and pre-training strategies of 16 recent noteworthy medical VLMs, and comprehensive discussion on evaluation metrics for assessing VLMs' performance in medical report generation and VQA. We also highlight current challenges facing medical VLM development, including limited data availability, concerns with data privacy, and lack of proper evaluation metrics, among others, while also proposing future directions to address these obstacles. Overall, our review summarizes the recent progress in developing VLMs to harness multimodal medical data for improved healthcare applications.
more »
« less
This content will become publicly available on July 1, 2026
GlyphPattern: An Abstract Pattern Recognition Benchmark for Vision-Language Models
Vision-Language Models (VLMs) have made rapid progress in reasoning across visual and textual data. While VLMs perform well on vision tasks that they are trained on, our results highlight key challenges in abstract pattern recognition. We present GlyphPattern, a 954 item dataset that pairs 318 human-written descriptions of visual patterns from 40 writing systems with three visual presentation styles.GlyphPattern evaluates abstract pattern recognition in VLMs, requiring models to understand and judge natural language descriptions of visual patterns. GlyphPattern patterns are drawn from a large-scale cognitive science investigation of human writing systems; as a result, they are rich in spatial reference and compositionality. Our experiments show that GlyphPattern is challenging for state-of-the-art VLMs (GPT-4o achieves only 55% accuracy), with marginal gains from few-shot prompting. Our detailed analysis reveals errors at multiple levels, including visual processing, natural language understanding, and pattern generalization.
more »
« less
- Award ID(s):
- 2326174
- PAR ID:
- 10624000
- Publisher / Repository:
- Association for Computational Linguistics
- Date Published:
- Volume:
- Findings of the Association for Computational Linguistics: ACL 2025
- Page Range / eLocation ID:
- 1140–1175
- Format(s):
- Medium: X
- Location:
- Vienna, Austria
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Vision-Language Models (VLMs) are trained on vast amounts of data captured by humans emulating our understanding of the world. However, known as visual illusions, human's perception of reality isn't always faithful to the physical world. This raises a key question: do VLMs have the similar kind of illusions as humans do, or do they faithfully learn to represent reality? To investigate this question, we build a dataset containing five types of visual illusions and formulate four tasks to examine visual illusions in state-of-the-art VLMs. Our findings have shown that although the overall alignment is low, larger models are closer to human perception and more susceptible to visual illusions. Our dataset and initial findings will promote a better understanding of visual illusions in humans and machines and provide a stepping stone for future computational models that can better align humans and machines in perceiving and communicating about the shared visual world.more » « less
-
Video-language models (VLMs), large models pre-trained on numerous but noisy video-text pairs from the internet, have revolutionized activity recognition through their remarkable generalization and open-vocabulary capabilities. While complex human activities are often hierarchical and compositional, most existing tasks for evaluating VLMs focus only on high-level video understanding, making it difficult to accurately assess and interpret the ability of VLMs to understand complex and fine-grained human activities. Inspired by the recently proposed MOMA framework, we define activity graphs as a single universal representation of human activities that encompasses video understanding at the activity, sub10 activity, and atomic action level. We redefine activity parsing as the overarching task of activity graph generation, requiring understanding human activities across all three levels. To facilitate the evaluation of models on activity parsing, we introduce MOMA-LRG (Multi-Object Multi-Actor Language-Refined Graphs), a large dataset of complex human activities with activity graph annotations that can be readily transformed into natural language sentences. Lastly, we present a model-agnostic and lightweight approach to adapting and evaluating VLMs by incorporating structured knowledge from activity graphs into VLMs, addressing the individual limitations of language and graphical models. We demonstrate a strong performance on activity parsing and few-shot video classification, and our framework is intended to foster future research in the joint modeling of videos, graphs, and language.more » « less
-
Recent advances in Visual Language Models (VLMs) have significantly enhanced video analytics. VLMs capture complex visual and textual connections. While Convolutional Neural Networks (CNNs) excel in spatial pattern recognition, VLMs provide a global context, making them ideal for tasks like complex incidents and anomaly detection. However, VLMs are much more computationally intensive, posing challenges for large-scale and real-time applications. This paper introduces EdgeCloudAI, a scalable system integrating VLMs and CNNs through edge-cloud computing. Edge- CloudAI performs initial video processing (e.g., CNN) on edge devices and offloads deeper analysis (e.g., VLM) to the cloud, optimizing resource use and reducing latency. We have deployed EdgeCloudAI on the NSF COSMOS testbed in NYC. In this demo, we will demonstrate EdgeCloudAI’s performance in detecting user-defined incidents in real-time.more » « less
-
In order for robots to operate effectively in homes and workplaces, they must be able to manipulate the articulated objects common within environments built for and by humans. Kinematic models provide a concise representation of these objects that enable deliberate, generalizable manipulation policies. However, existing approaches to learning these models rely upon visual observations of an object's motion, and are subject to the effects of occlusions and feature sparsity. Natural language descriptions provide a flexible and efficient means by which humans can provide complementary information in a weakly supervised manner suitable for a variety of different interactions (e.g., demonstrations and remote manipulation). In this paper, we present a multimodal learning framework that incorporates both vision and language information acquired in situ to estimate the structure and parameters that define kinematic models of articulated objects. The visual signal takes the form of an RGB-D image stream that opportunistically captures object motion in an unprepared scene. Accompanying natural language descriptions of the motion constitute the linguistic signal. We model linguistic information using a probabilistic graphical model that grounds natural language descriptions to their referent kinematic motion. By exploiting the complementary nature of the vision and language observations, our method infers correct kinematic models for various multiple-part objects on which the previous state-of-the-art, visual-only system fails. We evaluate our multimodal learning framework on a dataset comprised of a variety of household objects, and demonstrate a 23% improvement in model accuracy over the vision-only baseline.more » « less
