Artificial intelligence applied to digestive endoscopy
Journal Title: Applied Medical Informatics - Year 2019, Vol 41, Issue 0
Abstract
Introduction: In recent years, deep learning methods have improved significantly and have beenimplemented in fields such as medical imaging. Applying these techniques to digestiveendoscopy has led diagnosis rates for entities such as polyps similar or even better than humans.Materials and methods: We trained a convolutional neural network to classify medical images intotwo categories – with polyps or with normal mucosa – using about 800 images. For scalabilityand accessibility reasons, the architecture was implemented into a web interface. To ourknowledge, this is the first solution to emphasize the importance of scalability and accessibility.We developed an interface that can be used in real life scenarios and is easy to use, being webenabled and accessible from any device. Results: Experimental results show that our solution isfeasible and can be implemented in clinical practice. The model was evaluated on the test setand under these circumstances the final test accuracy was 100%. One limitation is the numberof images used for training. Whereas 800 images were used in total for training, only 100contained normal mucosa and 700 contained polyps. With future research, the number ofimages used will be increased and data enhancement techniques will be used, alongside withendoscopy videos. Conclusion: In conclusion, deep learning advances can be successfully appliedto biomedical fields such as digestive endoscopy for tasks such as polyp classification, with greatpotential of developing tools for medical professionals.
Authors and Affiliations
Andrei IOANOVICI, Sergiu CHERECHEȘ, Ștefan MĂRUȘTERI
Keywords
Related Articles
- EP ID EP655051
- DOI -
- Views 83
- Downloads 0
Data Sources and Information to be Collected for a Romanian Diabetes Register
Diabetes is a serious condition that damages many systems and organs, mainly the blood vessels and the nerves. The importance of the diabetes and its complications, along with the prevalence of the disease makes it a maj...
Chest X-Rays Image Classification in Medical Image Analysis
Chest X-Rays image classification is an active research area in medical image analysis as well as computer-aided diagnosis for radiology. The main goal is to improve the quality and productivity of radiologists’ task by...
Design of a Management Dashboard for the Intensive Care Unit: Determining Key Performance Indicators and their Required Capabilities
Introduction: Because the Intensive Care Unit (ICU) provides vital services to save the lives of critically ill patients, this unit has a special standing in health care systems; therefore it must be able to continuously...
Assessment of the Relationship between Hyperlipoproteinemias and Biliary Lithiasis
Aim: The purpose of this study was to bring new data regarding the prevalence of biliary lithiasis in Cluj district in regards to the role of hyperlipoproteinemias as a risk factor for this pathology. Material and Method...
Renal Tumor Cryoablation Planning. The Efficiency of Simulation on Reconstructed 3D CT Scan
[i]Introduction & Objective[/i]: Nephron-sparing surgical techniques risks are related to tumor relationships with adjacent anatomic structures. Complexity of the renal anatomy drives the interest to develop tools fo...