Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Aug;26(8):3950-3965.
doi: 10.1109/JBHI.2022.3160098. Epub 2022 Aug 11.

Artificial Intelligence for Colonoscopy: Past, Present, and Future

Artificial Intelligence for Colonoscopy: Past, Present, and Future

Wallapak Tavanapong et al. IEEE J Biomed Health Inform. 2022 Aug.

Abstract

During the past decades, many automated image analysis methods have been developed for colonoscopy. Real-time implementation of the most promising methods during colonoscopy has been tested in clinical trials, including several recent multi-center studies. All trials have shown results that may contribute to prevention of colorectal cancer. We summarize the past and present development of colonoscopy video analysis methods, focusing on two categories of artificial intelligence (AI) technologies used in clinical trials. These are (1) analysis and feedback for improving colonoscopy quality and (2) detection of abnormalities. Our survey includes methods that use traditional machine learning algorithms on carefully designed hand-crafted features as well as recent deep-learning methods. Lastly, we present the gap between current state-of-the-art technology and desirable clinical features and conclude with future directions of endoscopic AI technology development that will bridge the current gap.

PubMed Disclaimer

Figures

Fig. 1.
Fig. 1.
Diagram showing the colon anatomy.
Fig. 2.
Fig. 2.
Overview of the topics (in Sections II–V) summarized in this survey.
Fig. 3.
Fig. 3.
Examples (a) wall view; (b) lumen view; (c) spiral score and feedback; (d) retroflexion for viewing a difficult-to-reach area.
Fig. 4.
Fig. 4.
Examples of pixel-based interpretation on polyp images of the Kvasir V2 public dataset [17]: (a) input image, (b) gradient, (c) LRP, (d) Deep Taylor, and (e) Grad-CAM.
Fig. 5.
Fig. 5.
Examples of (a) ProtoPNet [187] and (b) two-level hierarchical concept-based interpretation [188] on polyp images of the Kvasir V2 public dataset [17]. The blue and yellow boxes indicate the image object that is used for prediction. Level 2 shows the object level concepts (e.g., the polyp object). Level 1 shows the low-level concepts (different shades of red colors and texture) that make up the polyp object. Thicker connecting lines indicate stronger influence of the lower-level to higher-level concepts.

Similar articles

Cited by

References

    1. World Health Organization International Agency for Research on Cancer, “Colorectal cancer source: Globocan 2020,” [Online]. Available: https://gco.iarc.fr/today/data/factsheets/cancers/10_8_9-Colorectum-fact...
    1. American Cancer Society, “Colorectum cancer statistics,” 2020. [Online]. Available: https://cancerstatisticscenter.cancer.org/?_ga=2.95264916.902125337.1581...
    1. Münzer B, Schoeffmann K, and Böszörmenyi L, “Content-based processing and analysis of endoscopic images and videos: A survey,” Multimedia Tools Appl, vol. 77, no. 1, pp. 1323–1362, 2018.
    1. Sánchez-Peralta LF, Bote-Curiel L, Picón A, Sánchez-Margallo FM, and Pagador JB, “Deep learning to find colorectal polyps in colonoscopy: A systematic literature review,” Artif. Intell. Med, vol. 108, pp. 1019–1023, 2020. - PubMed
    1. Prasath V, “Polyp detection and segmentation from video capsule endoscopy: A review,” J. Imag, vol. 3, no. 1, p. 1, 2016.

Publication types