Project Details
Abstract
The digital transformation in Pathology has opened the door to new opportunities to reshape this field. In the last decade, AI has made major advances in the detection, classification and prognostication of different cancer types in adult tissues based on Whole Slide Images (WSI). However, there has been limited effort targeting the development of AI models for pediatrics tissues. Accurate and fast diagnosis of pediatric diseases in the pathology department can positively impact the long-term population health and reduce the cost burden on the healthcare system. This grant is in partnership with Sidra Medicine to develop AI models for Pediatrics Pathology. The development of these models requires the creation of large-scale datasets of annotated images. Most of the existing ones focus on cell segmentation for a limited number of cell types. Such datasets are costly and time-consuming to create, as the task of annotating the complete cell boundaries is very tedious. We argue that adopting an annotation strategy of one point per cell, known as dot annotation, can be sufficient in several downstream tasks such as cell counting, detection, density estimation and region-growing segmentation. This project is in partnership with Sidra Medicine and Nvidia.
Submitting Institute Name
Hamad Bin Khalifa University (HBKU)
Sponsor's Award Number | HBKU-INT-VPR-TG-02-04 |
---|---|
Proposal ID | HBKU-OVPR-TG-Cycle-2-19 |
Status | Active |
Effective start/end date | 1/06/23 → 31/05/25 |
Collaborative partners
- Hamad Bin Khalifa University (lead)
- NVIDIA
- Sidra Medicine
Primary Theme
- Artificial Intelligence
Primary Subtheme
- AI - Healthcare
Secondary Theme
- Artificial Intelligence
Secondary Subtheme
- AI - Analytics & Decision Support
Keywords
- None
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.