Wireless Capsule Endoscopy Gets a Quantum boost for More Accurate Disease Detection
Doctors are increasingly using wireless Capsule endoscopy (WCE) – a non-invasive method – to diagnose digestive tract diseases. Though, analyzing the thousands of high-resolution images WCE produces is challenging for both humans adn traditional Artificial Intelligence (AI) due to the immense computing power required and the complexity of the data.
New research introduces a solution: Fused Quantum Dual-Backbone Network (FQDN). This innovative system combines the strengths of classical computing and quantum technology. It utilizes two established AI models (MobileNetV2 and EfficientNetB0) working together to extract detailed visual features from endoscopic images.
The key advancement lies in integrating quantum circuits into the analysis. Recognizing the limitations of current quantum hardware, researchers designed a specialized 4-qubit quantum circuit. This circuit processes complex disease patterns by converting image features into quantum data, leveraging the principles of superposition and entanglement – ultimately improving accuracy, especially for guiding biopsies.
The results are notable. FQDN dramatically reduces the model’s complexity, decreasing the number of parameters needed by up to 94.44% compared to traditional AI models. This makes it faster and more efficient without sacrificing accuracy.
In tests on a colorectal disease dataset, FQDN achieved 95.42% accuracy, surpassing established classical methods like VGG16 and InceptionV3. It particularly excelled at distinguishing between easily confused conditions like ulcers and polyps, and accurately identifying healthy tissue and esophagitis.
This hybrid approach demonstrates that combining classical and quantum computing can lead to more precise and efficient medical diagnostics,paving the way for wider request of Quantum Machine Learning in healthcare and faster,more accurate diagnoses for patients with gastrointestinal issues.
Source: Marzoug, N., et al. (2025). Quantum-Enhanced Dual-Backbone Architecture for Accurate Gastrointestinal disease Detection Using Endoscopic Imaging. BioMedInformatics, 5(3), 51.
Related reading