Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jun 27, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

SpatioPrompt: Learning Spatial Attention and Dynamic Prompts for Few-Shot Medical Image Anomaly Detection.

Liqiang Song, Yu Zhu, Junli Zhao

IEEE Journal of Biomedical and Health Informatics

|June 24, 2026

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Enhancement growth, water use efficiency and economic benefit for maize by drip irrigation in Northwest China.

Scientific reports·2023

Same author

DNA controllable peroxidase-like activity of Ti<sub>3</sub>C<sub>2</sub> nanosheets for colorimetric detection of microcystin-LR.

Analytical and bioanalytical chemistry·2023

Same author

pH regulates peptide-receptor perception.

Trends in plant science·2023

Same author

Identification and functional characterization of AcMYB113 in anthocyanin metabolism of Aesculus chinensis Bunge var. chinensis leaves.

Plant physiology and biochemistry : PPB·2023

Same author

Effect of dual residual risk of cholesterol and inflammation on all-cause mortality in patients with cardiovascular disease.

Cardiovascular diabetology·2023

Same author

The correlation of sacral table angle to spinopelvic sagittal alignment in healthy adults.

Journal of orthopaedic surgery and research·2023

Same journal

An Ultra-Lightweight Cross-scale Attention Mamba Network for Accurate Skin Lesion Segmentation.

IEEE journal of biomedical and health informatics·2026

Same journal

Explanation-Guided Reconstruction of Missing Clinical Features for Survival Prediction in Pancreatic Cancer.

IEEE journal of biomedical and health informatics·2026

Same journal

stDGCN: A dual-augmentation graph convolutional network for identifying spatial domains with attention mechanism.

IEEE journal of biomedical and health informatics·2026

Same journal

Patient-specific Biomechanical Investigation of Percutaneous Pulmonary Valves: Towards the Integration of Routinely Acquired Clinical Data and Fluid-structure Interaction Simulations.

IEEE journal of biomedical and health informatics·2026

Same journal

Cross-subject fMRI-to-Image with Visual-cortex 2D Representation and Pre-Training.

IEEE journal of biomedical and health informatics·2026

Same journal

PGCASurv: A Prior-Guided Cross-Attention Framework for Dynamic Survival Model with Longitudinal Data.

IEEE journal of biomedical and health informatics·2026

See all related articles

SpatioPrompt enhances few-shot medical anomaly detection by using spatial attention and dynamic prompts. This method improves performance on various imaging modalities without needing abnormal training data.

Area of Science:

Medical Imaging
Artificial Intelligence
Computer Vision

Background:

Few-shot medical image anomaly detection is challenging due to limited abnormal samples.
Existing vision-language models (VLMs) like MediCLIP struggle with fine-grained spatial features and prompt adaptivity.

Purpose of the Study:

To introduce SpatioPrompt, a parameter-efficient VLM adaptation framework for normal-only few-shot medical anomaly detection.
To improve the capture of spatial features and dynamic prompt generation for heterogeneous lesion patterns.

Main Methods:

Incorporated a spatial attention mechanism (CBAM-inspired) to focus on lesion regions and reduce background noise.
Developed a FewShotEnhancedRWKV decoder for dynamic, image-conditioned prompt generation using GRU-style gating and RWKV recurrence.

Related Experiment Videos

Last Updated: Jun 27, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

Main Results:

SpatioPrompt demonstrated significant improvements over MediCLIP across three modalities.
Achieved a 6.6% AUROC gain on BrainMRI (99.9% vs. 93.3% at 8-shot) and a 3.9% gain on BUSI (92.0% vs. 88.1% at 4-shot).
Showed consistent gains on CheXpert (73.8% vs. 70.7% at 16/32-shot).

Conclusions:

Combining spatial attention with dynamic prompt generation effectively enhances normal-only few-shot medical anomaly detection.
SpatioPrompt offers a promising approach for adapting VLMs to medical imaging tasks with limited data.