Researchers are revolutionizing how intestinal diseases are detected with the development of a tiny capsule camera, as per research conducted by scientists including Pål Anders Floor and Anuja Vats.
This technology, as discussed in their publications in IEEE Access and Scientific Reports respectively, offers a more comfortable and less invasive alternative to traditional methods like colonoscopies and gastroscopies.
Despite the promise of this technology, significant challenges have hindered its widespread adoption for over two decades.
The capsule camera travels through the digestive system for approximately eight hours, capturing over 50,000 images. However, analyzing this vast amount of data poses a significant challenge.
Pål Anders Floor, from the Department of Computer Science at NTNU in Gjøvik, has been researching creating a 3D representation of the intestines using capsule cameras.
Combined with machine learning, this approach could rapidly direct specialists’ attention to potential diseases, thereby streamlining the diagnostic process.
One of the primary issues with the current capsule technology is its passive movement through the intestines, leading to inconsistent imaging quality and potential missed diagnoses.
The capsule’s movement is uncontrolled, akin to the way food travels through the digestive system, resulting in varied picture quality and sometimes only capturing a single image of a diseased area. This is where machine learning comes in, as envisioned by Floor.
Machine learning has shown promise in disease detection, but it faces challenges like distinguishing between different diseases and the lack of sufficient data for training algorithms.
Anuja Vats, also from NTNU, is addressing these challenges by exploring how machine learning can enhance the capsule camera’s disease detection capabilities, despite limited data availability. She has been investigating the use of artificial data for training algorithms.
The team has created artificial images of various stages of intestinal diseases using a machine learning framework developed by Nvidia.
These images, validated by gastrointestinal specialists for their realism, could potentially be used to develop comprehensive training programs for capsule endoscopy.
Floor and Vats’ research paths, though different, complement each other. Combining disease-identifying machine learning techniques with 3D intestinal models could lead to a tool not only for improved capsule camera examinations but also for surgical planning and practice.
This development is especially relevant considering the variability in the anatomy of the digestive tract among individuals, as suggested by a recent study from North Carolina State University.
Despite the progress, both scientists acknowledge that more work is needed before capsule cameras can fully replace traditional examinations.
However, their primary goal is to provide a comfortable and accessible alternative for people hesitant to undergo more invasive procedures.
This technology holds the potential to significantly improve the comfort and efficiency of intestinal health diagnoses, potentially benefiting countless individuals reluctant to undergo traditional examinations.
If you care about colon health, please read studies about whether aspirin could lower colon cancer risk in older people, and this drug may lower death risk in colon cancer.
For more information about colon health, please see recent studies about how to protect yourself from colon cancer, and results showing this vitamin level in the body is linked to colon cancer risk.
The research findings can be found in IEEE Access.
Copyright © 2023 Knowridge Science Report. All rights reserved.
