Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 6, 2026

A Novel Dual-Modal Deep Learning Approach for Real-Time Removal of Hepatic Fluorescence in Indocyanine Green-Guided Laparoscopic Cholecystectomy
09:21

A Novel Dual-Modal Deep Learning Approach for Real-Time Removal of Hepatic Fluorescence in Indocyanine Green-Guided Laparoscopic Cholecystectomy

Published on: April 17, 2026

44

USSGAN: Unsupervised Spectral and Spatial Attention-Based Generative Adversarial Network for Cholangiocarcinoma

Sikhakolli Sravan Kumar1, Anuj Deshpande1, Pooja A Nair1

  • 1Electronics and Communication Department, SRM University-AP, Mangalagiri, Andhra Pradesh 522240, India.

Chemical & Biomedical Imaging
|December 26, 2025
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Metasurface-enabled high-gain circularly polarized antenna array for sensing in dust-laden environments.

Scientific reports·2026
Same author

Quad-port mm-wave MIMO antenna with a unified ground plane for 5G and beyond applications.

Scientific reports·2026
Same author

Advanced receiver design for AF-FD cooperative schemes.

Scientific reports·2026
Same author

Energy-efficiency optimization and comparison for IRS-assisted bidirectional relay and direct transmissions.

Scientific reports·2026
Same author

Comparative analysis of deep learning and traditional methods for IoT botnet detection using a multi-model framework across diverse datasets.

Scientific reports·2025
Same author

Overview of photon-counted three-dimensional imaging and related applications.

Optics express·2025

This study introduces Unsupervised Spectral and Spatial Attention-based GAN (USSGAN) for detecting liver bile duct cancer (cholangiocarcinoma). This AI approach accurately identifies cancerous regions using hyperspectral imaging without needing labeled data, offering a fast and efficient diagnostic tool.

Area of Science:

  • Medical Imaging
  • Artificial Intelligence
  • Oncology

Background:

  • Cholangiocarcinoma (bile duct cancer) has a low survival rate, making early detection crucial.
  • Conventional imaging (CT, MRI) has limitations; Hyperspectral Imaging (HSI) offers a non-invasive alternative.
  • Supervised learning for HSI analysis requires extensive annotated datasets, which are difficult to acquire.

Purpose of the Study:

  • To develop an unsupervised learning method for cholangiocarcinoma detection using Hyperspectral Imaging (HSI).
  • To improve the accuracy and efficiency of cancer diagnosis by leveraging Generative Adversarial Networks (GANs).

Main Methods:

  • Proposed an Unsupervised Spectral and Spatial Attention-based GAN (USSGAN) for classifying and segmenting cancerous regions.
  • Integrated adaptive step size into Tasmanian Devil Optimization (TDO), creating Enhanced Tasmanian Devil Optimization (ETDO), to improve convergence and feature capture.
Keywords:
cholangiocarcinomaenhanced tasmanian devil optimizationgenerative adversarial networkshyperspectral imagingunsupervised learning

Related Experiment Videos

Last Updated: May 6, 2026

A Novel Dual-Modal Deep Learning Approach for Real-Time Removal of Hepatic Fluorescence in Indocyanine Green-Guided Laparoscopic Cholecystectomy
09:21

A Novel Dual-Modal Deep Learning Approach for Real-Time Removal of Hepatic Fluorescence in Indocyanine Green-Guided Laparoscopic Cholecystectomy

Published on: April 17, 2026

44
  • Utilized a publicly available multidimensional choledochal cholangiocarcinoma dataset for testing.

    • Main Results:

    • USSGAN achieved superior performance compared to existing methods on the cholangiocarcinoma dataset.
    • Demonstrated high accuracy with an overall accuracy (OA) of 98.03%, average accuracy (AA), and Cohen's Kappa.
    • Ablation studies validated the effectiveness of the proposed enhancements (ETDO and attention mechanisms).

    Conclusions:

    • USSGAN provides a robust, computationally efficient, and accurate unsupervised method for cholangiocarcinoma detection via HSI.
    • The lightweight model enables real-time clinical deployment, delivering results within a minute.
    • Achieved performance comparable to expert pathologists, offering a practical solution for early cancer diagnosis.