Communication Dans Un Congrès Année : 2023

Ex Vivo Breast Cancer Tissue Classification Using Hyperspectral Endoscopy Imaging with Deep Learning

Résumé

Current endoscopy examination practices rely on the doctor's experience and the pathologist's report to make a diagnosis which is labor-intensive and time-consuming. In this paper, we developed a hyperspectral imaging (HSI) endoscopy imaging system that can diagnose breast cancer tissue samples to improve operating room efficiency and patients' clinical experience. Hyperspectral imaging is a technique that captures both spatial (x, y) and spectral (λ) information reflecting the morphological and functional features of a sample. A 3D convolutional neural network deep learning model was developed to classify ductal carcinoma, non-ductal carcinoma, and normal human breast tissue ex vivo based on the hyperspectral imaging datacubes. The result showed that our system could perform the classification and make diagnosis predictions rapidly and accurately.

Fichier non déposé

Dates et versions

hal-04294347 , version 1 (19-11-2023)

Identifiants

Citer

Shuyan Zhang, Shuwen Zeng, Ryan Ron Zee Tan, Wenjun Liao, Malini Olivo. Ex Vivo Breast Cancer Tissue Classification Using Hyperspectral Endoscopy Imaging with Deep Learning. 2023 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC), Jun 2023, Munich, Germany. pp.1-2, ⟨10.1109/CLEO/Europe-EQEC57999.2023.10232591⟩. ⟨hal-04294347⟩

Collections

52 Consultations
0 Téléchargements

Altmetric

Partager

  • More