DL for Accurate Histopathological Cancer Detection
Histopathological Cancer Detection Using Intra-Domain Transfer Learning and Ensemble Learning
Tags: Universiti Malaya, Malaysia, Healthcare & Lifesciences
This technology uses intra-domain transfer learning and ensemble learning to improve histopathological cancer detection. The deep learning models developed are highly accurate, achieving 99.78% accuracy on the GasHisSDB dataset, 85.69% on Chaoyang, and 99.17% on CPTAC-CCRCC. These models can automate the labor-intensive task of image screening, assisting pathologists and reducing variability in diagnoses. Applications include early cancer detection to reduce mortality rates. The model performs well on low-resolution images, making it effective for real-world medical settings.
IP Type or Form Factor: Discovery & Research; Software & Algorithm
TRL: 5 - prototype ready for testing in intended environment
Industry or Tech Area: Diagnostics & Screening; Healthcare Provider
