Related Experiment Video
Updated: Dec 15, 2025
10:39
A Label-Free Segmentation Approach for Intravital Imaging of Mammary Tumor Microenvironment
Published on: May 24, 2022
2.7K
Nerve Segmentation with Deep Learning from Label-Free Endoscopic Images Obtained Using Coherent Anti-Stokes Raman
Naoki Yamato1, Mana Matsuya1, Hirohiko Niioka2
1Graduate School/Faculty of Information Science and Technology, Hokkaido University, Sapporo 060-0814, Japan.
Biomolecules
|July 12, 2020
Summary
Deep learning accurately segments nerves from label-free endoscopic images using Coherent Anti-Stokes Raman Scattering (CARS) for safer surgery. Pre-training on fluorescence images significantly improved nerve segmentation accuracy and performance.
Area of Science:
- Medical Imaging
- Surgical Technology
- Artificial Intelligence in Medicine
Background:
- Nerve-sparing surgery requires precise identification of peripheral nerves.
- Coherent Anti-Stokes Raman Scattering (CARS) endoscopy visualizes myelinated nerves via lipid signals in a label-free manner.
- Accurate nerve segmentation is crucial as lipid signals are present in non-nerve tissues.
Purpose of the Study:
- To develop and evaluate a deep learning model for semantic segmentation of nerves from label-free CARS endoscopic images.
- To enhance nerve identification for improved outcomes in nerve-sparing surgery.
Main Methods:
- A U-Net deep learning model with a VGG16 encoder was utilized.
- The model was pre-trained on fluorescence images and fine-tuned on a small dataset of CARS endoscopy images.
- Nerve segmentation was performed using 24 CARS and 1,818 fluorescence nerve images from rabbit prostates.
Main Results:
- The deep learning model achieved label-free nerve segmentation with a mean accuracy of 0.962 and an F1 score of 0.860.
- Pre-training on fluorescence images significantly improved segmentation performance (p < 0.05).
Conclusions:
- Semantic segmentation of label-free endoscopic images enables precise nerve identification.
- This technique promises safer endoscopic surgeries, reduced patient dysfunction, and improved post-operative prognosis.
