Infrared target detection improves with zoom guided interactive method
IRSTD-Agent: Agentic Infrared Small Target Detection via Zoom-Guided Interaction Learning
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Detecting tiny objects in infrared images is important for things like monitoring the ocean or skies. The authors created a system called IRSTD-Agent that helps a computer model decide where to look and how closely to examine parts of an image. This step-by-step search uses different tools to find and pinpoint small targets more accurately. Their method teaches the model to choose the right tools based on example searches and improves results on standard infrared image tests.
What this means in practice
- •For maritime surveillance operators: Improve the detection and precise location of small infrared targets in maritime surveillance images using adaptive zoom-based search.
- •For aerial surveillance teams: Enable enhanced identification and detailed localization of small infrared objects in aerial monitoring systems through interactive agent methods.
Authors
Jiawen Xi, Yu Zhang, Tianyi Zhao, Zhu Liu, Maoxun Yuan, Xingxing Wei
Abstract
Infrared small-target detection plays an important role in maritime monitoring and aerial surveillance. Although multimodal large language models (MLLMs) offer promising capabilities for visual understanding, existing MLLM-based approaches struggle to precisely localize infrared small targets. In this paper, we propose IRSTD-Agent, an agentic framework for infrared small target detection through dynamic visual search. The framework enables an MLLM to adaptively determine where and at what scale to inspect an image and progressively gather fine-grained visual evidence for precise target localization. Five complementary visual tools (PROPOSAL, ZOOM, DETECT, DROP and REFINE) support object candidate discovery, adaptive observation, target localization, hypothesis rejection, and target extent refinement, together enabling a coordinated search process over original-resolution images. To teach the MLLMs to conduct this search, we introduce Zoom-guided Interaction Learning, which uses annotation-derived interaction trajectories to supervise tool selection and the corresponding arguments. Through extensive experiments on WideIRSTD-Full and IRSTD-1k datasets, we demonstrate that IRSTD-Agent outperforms the evaluated vision-language models and enhances the precise localization capabilities of MLLMs in IRSTD tasks.