* If you want to update the article please login/register
Capsule Endoscopy pioneered the study of the small intestine, overcoming the limitations of traditional endoscopy. Nonetheless, reviewing CE images is time-consuming. Convolutional Neural Networks (CNNs) are an artificial intelligence framework with high success rates for image processing. In CE images, protruding lesions of the small intestine display a wide range in morphologic variation. Methods: A CNN was created using a collection of CE images of protruding lesions or normal mucosa/other findings. These photos were embedded in a CNN timeline with transfer learning. We evaluated the network's results by determining its sensitivity, specificity, accuracy, positive predictive value, and negative predictive value. The results: A CNN was created based on a total of 21,320 CE photographs. With a 97. 1% success, the algorithm automatically detected small-bowel protruding lesions. Conclusion: We created an effective CNN for the automatic detection of enteric protruding lesions in a variety of morphologies.
Source link: https://doi.org/10.1016/j.gastha.2022.04.008
Vision Transformer is quickly emerging as a new leader in computer vision with its outstanding success in several tasks. On a small-scale imbalanced capsule endoscopic image database, it is impossible for us to use ViT to train from scratch. We started from scratch on two publicly available datasets for capsule endoscopy disease classification, with 79. 15% precision on the Kvasir-Capsule dataset and 98. 63% on the binary classification task of the Red Lesion Endoscopy dataset.
Source link: https://doi.org/10.3390/electronics11172747
People of various ages are enrolled in clinical examination, but no research specifically focused on the young. Patients with OGIB were divided into four types of lesions, and the possibility of bleeding among patients with OGIB were assessed based on Saurin classification. Uluration was the most commonest P2 lesions among patients with OGIB in which 134 patients had a total of 216 lesions, followed by angioectasia and telangiectasia. Patients with overt OGIB, completed rate, and diagnostic yield were all higher among patients with overt OGIB, and disease categories of overt OGIB were different from those for occult OGIB. Conclusion CE is an outstanding tool for detecting lesions in young adults, and it may play a role in determining the bleeding risk of young adults with OGIB.
Source link: https://doi.org/10.1002/jgh3.12801
In particular, we recommend using biological residual cues to specifically invite the network to collect pathology traces. While residuals are recovered using well-established 2D wavelet decomposition, we do also suggest using color channels to learn discriminative cues in WCE images. With a comprehensive benchmark for WCE abnormality and multi-class classification, we illustrate the generalizability of the suggested strategy on both datasets, where our findings are more reliable than the state-of-the-art with much fewer labels in abnormality sensitivity on several of nine pathologies and establish a new benchmark with specificity across classes.
Source link: https://doi.org/10.1109/ACCESS.2022.3201515
History: Colon capsule endoscopy is a method used by patients who are unable or against contraindications for traditional colonoscopy. Colorectal cancer screening may be able to benefit greatly from widespread use of a non-invasive technique such as CCE. In CCE photographs, we wanted to create an artificial intelligence algorithm based on a convolutional neural network architecture for automatic detection of colonic protruding lesions. Images of patients with colonic protruding lesions or patients with normal colonic mucosa or other pathologic findings were included in this database. The introduction of AI technologies to CCE may improve diagnostic accuracy and acceptance for the screening of colorectal neoplasia.
Source link: https://doi.org/10.3390/diagnostics12061445
To improve the accuracy of the reading process of wireless capsule endoscopy images or movies, computer-aided diagnosis has been used. However, there are no studies that objectively determine the results of CAD models in diagnosing gastrointestinal protruded lesions. Objective: Using wireless capsule endoscopic images, this review sought to determine the diagnostic results of CAD models for digestive protruded lesions. Methods: Core databases were searched for reports based on CAD models for the diagnosis of gastrointestinal protruded lesions using wireless capsule endoscopy, and diagnostic results were published. The pooled area under the curve, accuracy, specificity, and diagnostic odds ratio of CAD models for the diagnosis of protruded lesions were 0. 95, 0. 89, 0. 91, and 74, respectively, with 0. 95, 0. 91, 0. 91, and 74 for the diagnosis of protruded lesions. Conclusion: CAD models demonstrated excellent results for the optical diagnosis of gastrointestinal protruded lesions based on wireless capsule endoscopy.
Source link: https://doi.org/10.3390/jpm12040644
In addition, future multi-functioning robotic capsule endoscopy units may include advanced functions such as active system control over capsule motion, medical delivery systems, semi-surgical instruments, and biopsy. In this regard, the wireless power transmission device has received increasing attention from researchers trying to solve this problem. The current technological advancements have limited systematic reviews of WPT for WCE. The WPT system for this WCE application is still in the initial stage, and it should be noted that improvements regarding system efficiency, stability, and patient safety are all still in the process.
Source link: https://doi.org/10.3390/s140610929
It's a difficult challenge to detect abnormalities in wireless capsule endoscopy images. This study, based on state-of-the-art deep learning networks, suggests that an automated system for detecting and analyzing ulcers in WCE images. The medical image datasets used in this study were obtained from WCE video frames, which were obtained from WCE video frames. Both historic CNN architectures, namely the AlexNet and the GoogLeNet, are extensively reviewed in object classification into ulcer or non-ulcer. In addition, we review and analyze the photographs found as containing ulcer objects in order to determine the effectiveness of the used CNNs.
Source link: https://doi.org/10.3390/s19061265
A new low-cost, lossless image compression scheme for capsule endoscopy is introduced in this paper. Compressor with a low-cost YEF color space converter and variable-length predictive with a combination of Golomb-Rice and unary encoding, as well as Golomb-Rice and unary encoding. Unlike transform based algorithms, the compressor can be interfacing with commercial image sensors that send pixel data in a raster-scan manner, eliminating the need for having large buffer memory. The compression algorithm can be used with white light imaging and narrow band imaging with an average compression ratio of 78% and 84% respectively. On a single, low-power, 65-nm field programmable gate arrays chip, a complete capsule endoscopy system is constructed. The proposed algorithm provides a solution to wireless capsule endoscopy with lossless and yet acceptable compression, according to the study findings.
Source link: https://doi.org/10.3390/s141120779
* Please keep in mind that all text is summarized by machine, we do not bear any responsibility, and you should always check original source before taking any actions