Improved Accuracy in Optical Diagnosis of Colorectal Polyps Using Convolutional Neural Networks with Visual Explanations.

BACKGROUND & AIMS: Narrow-band imaging (NBI) can be used to determine whether colorectal polyps are adenomatous or hyperplastic. We investigated whether an artificial intelligence (AI) system can increase the accuracy of characterizations of polyps by endoscopists of different skill levels. METHODS: We developed convolutional neural networks (CNNs) for evaluation of diminutive colorectal polyps, based on efficient neural architecture searches via parameter sharing with augmentation using NBIs of diminutive (≤5 mm) polyps, collected from October 2015 through October 2017 at the Seoul National University Hospital, Healthcare System Gangnam Center (training set). We trained the CNN using images from 1100 adenomatous polyps and 1050 hyperplastic polyps from 1379 patients. We then tested the system using 300 images of 180 adenomatous polyps and 120 hyperplastic polyps, obtained from January 2018 to May 2019. We compared the accuracy of 22 endoscopists of different skill levels (7 novices, 4 experts, and 11 NBI-trained experts) vs the CNN in evaluation of images (adenomatous vs hyperplastic) from 180 adenomatous and 120 hyperplastic polyps. The endoscopists then evaluated the polyp images with knowledge of the CNN-processed results. We conducted mixed-effect logistic and linear regression analyses to determine the effects of AI assistance on the accuracy of analysis of diminutive colorectal polyps by endoscopists (primary outcome). RESULTS: The CNN distinguished adenomatous vs hyperplastic diminutive polyps with 86.7% accuracy, based on histologic analysis as the reference standard. Endoscopists distinguished adenomatous vs hyperplastic diminutive polyps with 82.5% overall accuracy (novices, 73.8% accuracy; experts, 83.8% accuracy; and NBI-trained experts, 87.6% accuracy). With knowledge of the CNN-processed results, the overall accuracy of the endoscopists increased to 88.5% (P < .05). With knowledge of the CNN-processed results, the accuracy of novice endoscopists increased to 85.6% (P < .05). The CNN-processed results significantly reduced endoscopist time of diagnosis (from 3.92 to 3.37 seconds per polyp, P = .042). CONCLUSIONS: We developed a CNN that significantly increases the accuracy of evaluation of diminutive colorectal polyps (as adenomatous vs hyperplastic) and reduces the time of diagnosis by endoscopists. This AI assistance system significantly increased the accuracy of analysis by novice endoscopists, who achieved near-expert levels of accuracy without extra training. The CNN assistance system can reduce the skill-level dependence of endoscopists and costs.

An Wearable Energy Expenditure Analysis System based on the 15-channel Whole-body Segment Acceleration Measurement.

The measurement of the amount of energy utilized during physical activity has generated considerable interests from various groups ranging from exercise physiologists to nutritionists and fitness center workers. To date, however, the existing energy expenditure estimation methods are not so reliable and compact. In this paper, we propose a new method for accurately and easily estimating energy expenditure during physical activity with a novel algorithm. This method involves acquiring acceleration signals through a 15-channel whole-body segment acceleration measurement system and then estimating the calories expended using a newly developed algorithm. The results of 3 subjects' experiments were compared with a commercially available mask type indirect calorimeter and a 9-axis accelerometry-based calorimeter. The results demonstrate that the proposed method provides a new and reliable way to estimate human energy expenditure during physical activity.

Artificial neural network-based estimation of the eyeball position using the magnetic contact lens sensing technique.

We have created an artificial neural network based approach for measuring eye movement using a magnetic contact lens sensing technique. The sensor array is based on using four magnetoresistive sensors. A two-layer feed-forward artificial neural network was used and an artificial eyeball model was made for the test. The neural network is trained with sample data obtained from nine spots. After training, we compared the position calculated from the developed system with the real one. The result shows that there is a good linear relationship between them. This indicates the developed system is capable of recording the position of the eyeball with a high degree of accuracy.

A wireless measurement system for three-dimensional ocular movement using the magnetic contact lens sensing technique.

A new and innovative method to measure the eye movement in a wireless manner was proposed. We verified the feasibility of our idea by fabrication and performance test of a prototype system. The prototype system consisted of a contact lens with a ring-shaped thin magnet, and eyeglasses frame-shaped PCB with analog/digital signal processing circuitry as well as four magnetoresistive sensors. This new method based on the magnetic contact lens sensing technique (MCLST) is expected to overcome all the disadvantages of the existing techniques.