Gilberto Ka Kit Leung

Title: Gilberto Ka Kit Leung: Medical Imaging and Data Segmentation

Introduction

Gilberto Ka Kit Leung is an inventor based in Hong Kong, China, whose patent work concerns computer-assisted analysis of medical images. He is associated with the University of Hong Kong.

Latest Patents

Leung is named on 1 patent: “Dissimilar-paired neural network architecture for data segmentation.” The patent describes a computer-implemented system for processing and analyzing non-contrast-enhanced computed tomography images. It combines two U-Net architectures with equal weights in a Siamese arrangement and a linked Dissimilar block, forming a Dissimilar-Siamese-U-Net architecture.

The system is designed to analyze separate images of the brain’s left and right hemispheres. Its potential diagnostic applications include acute ischemic stroke and thromboembolus, using detection of the hyperdense middle cerebral artery sign.

Career Highlights

The patent reflects Leung’s work at the intersection of neural-network architecture, medical imaging, and computer-assisted diagnosis. The disclosed approach applies paired image analysis to support the identification of clinically relevant signs in brain scans.

Collaborations

Leung’s listed coworkers include Philip Leung Ho Yu and Jia You. His association with the University of Hong Kong connects his work to an academic research environment.

Conclusion

Gilberto Ka Kit Leung’s patent describes a paired neural-network system for segmenting and analyzing brain imaging data, with a focus on supporting stroke-related diagnosis.