AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Text-IRSTD: Leveraging Semantic Text to Promote Infrared Small Target Detection in Complex Scenes

Feng Huang, Shuyuan Zheng, Zhaobing Qiu, Huanxian Liu, Huanxin Bai, Liqiong Chen

ICCV 2025 · Vision-language and multimodal › Human and embodied modeling · approach: Region-level vision-language alignment · 11 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add fuzzy semantic text prompt guidance to the IRSTD network (ablation)Text-IRSTD · IRSTD-1kIoU (%): 65.50 -> 69.57Improved
Add text guidance on NUDT-SIRST where visual features are already strong (ablation)Text-IRSTD · NUDT-SIRSTIoU (%): 95.25 -> 95.84Improved
Use fuzzy scenario prompt instead of conventional class prompt (ablation)Text-IRSTD · IRSTD-1kIoU (%): 62.22 -> 69.57Improved
Add scene background description to the prompt (ablation)Text-IRSTD · IRSTD-1kIoU (%): 62.22 -> 66.65Improved
Add TGFA block to baseline (ablation)Text-IRSTD · IRSTD-1kIoU (%): 61.67 -> 67.75Improved
Add TGSI block to baseline (ablation)Text-IRSTD · IRSTD-1kIoU (%): 61.67 -> 66.19Improved
Combine TGFA and TGSI (ablation)Text-IRSTD · IRSTD-1kIoU (%): 67.75 -> 69.57Improved
Add detail perception module (DPM) in TGFA (ablation)TGFA · IRSTD-1kIoU (%): 67.16 -> 69.57Improved
Add channel attention (CA) in TGFA (ablation)TGFA · IRSTD-1kIoU (%): 68.01 -> 69.57Improved
Modulate image features with text-derived gain and offset instead of concatenation (ablation)TGSI · IRSTD-1kIoU (%): 67.73 -> 69.57Improved
Add gain parameter alpha_t in TGSI (ablation)TGSI · IRSTD-1kIoU (%): 66.65 -> 69.57Improved
Add offset parameter beta_t in TGSI (ablation)TGSI · IRSTD-1kFa (1e-6): 16.22 -> 14.97No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy