Abstract
The present invention discloses a computer system for processing whole section images (WSI) to detect cancer (CA), including a machine learning model for detecting CA based on WSI, which can process the average of cells identified on the WSI. Cell characteristics. Each cell feature is a descriptive statistic for the cell, advantageously allowing the construction of training and test datasets without the expensive annotation process of labeling each cell pixel by pixel on WSI training samples. In addition to predicting CA or non-CA cases for WSI, this model can also be used to: identify suspected CA cases for priority assessment if predicted for WSI to be non-CA cases; generate tumor probability heatmaps of WSI for use in Visualization of potential CA areas on WSI to assist in pathological assessment; and in assessing quality control of sorting systems prior to implementation in clinical settings. The invention also discloses a sorting system for diagnosing CA cases.
| Original language | English |
|---|---|
| Patent number | US12,462,385 |
| Filing date | 8/06/23 |
| Publication status | Published - 2026 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 3 Good Health and Well-being
Fingerprint
Dive into the research topics of 'Weakly-Supervised System, Method and Workflow for Processing Whole Slide Image for Disease Detection'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver