Related Experiment Video
Updated: Sep 12, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
CLIF-Net: Intersection-guided Cross-view Fusion Network for Infection Detection from Cranial Ultrasound
Mingzhao Yu1, Mallory R Peterson2, Kathy Burgoine3
1Department of Electrical Engineering, Pennsylvania State University, University Park, PA 16802 USA.
Insights
Detecting serious bacterial infection in infants using cranial ultrasound (cUS) is crucial. A new AI framework, CLIF-Net, effectively analyzes multi-view cUS images for improved infant infection detection.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neonatal Health
Background:
- Early detection of possible serious bacterial infection (pSBI) in infants is critical for timely treatment and improved outcomes.
- Cranial ultrasound (cUS) is a valuable imaging modality for neonatal assessment, providing multi-view data (coronal and sagittal).
- Existing methods for pSBI detection using cUS may not fully exploit the rich information available from multi-view imaging.
Purpose of the Study:
- To develop and evaluate a novel deep learning framework, CLIF-Net, for enhanced detection of pSBI in infants using multi-view cUS images.
- To leverage the geometric overlap between coronal and sagittal cUS views to create a robust 3D representation for improved diagnostic accuracy.
- To improve upon state-of-the-art methods for identifying serious bacterial infections in newborns.
Main Methods:
- Development of the intersection-guided Cross-view Local- and Image-level Fusion Network (CLIF-Net), a deep learning framework utilizing two distinct convolutional neural network branches.
- Implementation of multi-level fusion blocks with cross-attention modules to extract and enhance semantic features from intersecting regions of coronal and sagittal cUS images.
- Integration of enhanced features through an image-level fusion layer to output class probabilities for pSBI and non-pSBI.
Main Results:
- CLIF-Net demonstrated substantially enhanced performance in detecting pSBI compared to existing techniques.
- The framework effectively exploited multi-view cUS imagery, creating a robust 3D representation for pSBI detection.
- Evaluation on a dataset of 302 cUS scans from Uganda showed superior results, surpassing current state-of-the-art infection detection methods.
Conclusions:
- The CLIF-Net framework offers a novel and effective approach for pSBI detection in infants by utilizing multi-view cUS data.
- Exploiting the geometric relationship and cross-view features significantly improves the accuracy of infection detection in newborns.
- This method represents a significant advancement in leveraging AI for neonatal infection diagnosis, with potential for broader clinical application.
Abstract:
This paper addresses the problem of detecting possible serious bacterial infection (pSBI) of infancy, i.e. a clinical presentation consistent with bacterial sepsis in newborn infants using cranial ultrasound (cUS) images. The captured image set for each patient enables multi-view imagery: coronal and sagittal, with geometric overlap. To exploit this geometric relation, we develop a new learning framework, called the intersection-guided Cross-view Local- and Image-level Fusion Network (CLIF-Net). Our technique employs two distinct convolutional neural network branches to extract features from coronal and sagittal images with newly developed multi-level fusion blocks. Specifically, we leverage the spatial position of these images to locate the intersecting region. We then identify and enhance the semantic features from this region across multiple levels using cross-attention modules, facilitating the acquisition of mutually beneficial and more representative features from both views. The final enhanced features from the two views are then integrated and projected through the image-level fusion layer, outputting pSBI and non-pSBI class probabilities. We contend that our method of exploiting multi-view cUS images enables a first of its kind, robust 3D representation tailored for pSBI detection. When evaluated on a dataset of 302 cUS scans from Mbale Regional Referral Hospital in Uganda, CLIF-Net demonstrates substantially enhanced performance, surpassing the prevailing state-of-the-art infection detection techniques.
