Transfer Learning with Human Corneal Tissues: An Analysis of Optimal Cut-Off Layer

2018/06/19 by Nadezhda Prodanova, Johannes Stegmaier, Prodanova, Nadezhda +13
Computer Science · Medicine · #AI in cancer detection #Corneal surgery and disorders #FOS: Computer and information sciences #Neural and Evolutionary Computing (cs.NE) #Retinal Imaging and Analysis

paper · pdf · doi:10.48550/arxiv.1806.07073

openalex publication_date 2018/06/19 · openalex created_date 2018/06/29 · openalex updated_date 2026/07/28

Abstract

Transfer learning is a powerful tool to adapt trained neural networks to new tasks. Depending on the similarity of the original task to the new task, the selection of the cut-off layer is critical. For medical applications like tissue classification, the last layers of an object classification network might not be optimal. We found that on real data of human corneal tissues the best feature representation can be found in the middle layers of the Inception-v3 and in the rear layers of the VGG-19 architecture.

Citations

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• ImageNet Large Scale Visual Recognition Challenge
• What makes ImageNet good for transfer learning?

Related

• Deep Learning for Cornea Microscopy Blind Deblurring
• Inspect Transfer Learning Architecture with Dilated Convolution
• Learning to Segment Corneal Tissue Interfaces in OCT Images
• Channel Scaling: A Scale-and-Select Approach for Transfer Learning
• Towards Optimal Convolutional Transfer Learning Architectures for Breast Lesion Classification and ACL Tear Detection