Improvement of Colon Polyp Detection Performance by Modifying the Multi-scale Network Structure and Data Augmentation
Journal of Electrical Engineering and Technology, ISSN: 2093-7423, Vol: 17, Issue: 5, Page: 3057-3065
2022
- 6Citations
- 4Captures
Metric Options: Counts1 Year3 YearSelecting the 1-year or 3-year option will change the metrics count to percentiles, illustrating how an article or review compares to other articles or reviews within the selected time period in the same journal. Selecting the 1-year option compares the metrics against other articles/reviews that were also published in the same calendar year. Selecting the 3-year option compares the metrics against other articles/reviews that were also published in the same calendar year plus the two years prior.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Example: if you select the 1-year option for an article published in 2019 and a metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019. If you select the 3-year option for the same article published in 2019 and the metric category shows 90%, that means that the article or review is performing better than 90% of the other articles/reviews published in that journal in 2019, 2018 and 2017.
Citation Benchmarking is provided by Scopus and SciVal and is different from the metrics context provided by PlumX Metrics.
Article Description
This study proposed a computer-assisted diagnosis system that detects polyps during colonoscopy using a multiscale network structure. Medical data require institutional review board approval, and collecting sufficient data is challenging for several reasons. The amount of data may be small thereby resulting in overfitting. This study attempted to increase the amount of data available to solve this problem. Autoaugment and the policy applied to the CIFAR-10 dataset were used. This data augmentation can be learned immediately without review by a colonist because no changes in the shape of the polyp occur during colonoscopy with minimal movement in location. The object detection network used was YOLOv4, which is capable of multiscale learning. Multiscale learning is advantageous in detecting an object regardless of the size of the lesion because it can extract features of various sizes through one learning. In this study, the learning advantages of multiple scales were reinforced via the addition of scales to YOLOv4, while the learning accuracy was improved by changing the activation function. Therefore, the changed activation function can continuously extract features when updating the layer weight. When using all the methods presented, mAP exhibited the highest performance at 98.36.
Bibliographic Details
Springer Science and Business Media LLC
Provide Feedback
Have ideas for a new metric? Would you like to see something else here?Let us know