Evaluation method of matching points for improving accuracy of position estimation using monocular endoscope

Fumiya Okuno; Kayo Yoshimoto; Hideya Takahashi; Goro Hamagishi

doi:10.1117/12.3001311

12 March 2024 Evaluation method of matching points for improving accuracy of position estimation using monocular endoscope

Fumiya Okuno, Kayo Yoshimoto, Hideya Takahashi, Goro Hamagishi

Author Affiliations +

Proceedings Volume 12820, Endoscopic Microscopy XIX; 128200A (2024) https://doi.org/10.1117/12.3001311
Event: SPIE BiOS, 2024, San Francisco, California, United States

Conference Poster

Abstract

Endoscopic examination has the advantage of being minimally invasive, but the field of view is narrow and the shape of the gastrointestinal tract creates blind spots that may cause lesions to be overlooked. It is thought that these problems can be solved by performing 3D reconstruction from images using Visual SLAM (Simultaneous Localization and Mapping) and grasping the structure in three dimensions. However, when this method, which requires matching between images, is applied inside the gastrointestinal tract, the similar environment in the gastrointestinal tract continues repeatedly, which prevents appropriate matching and appears as noise during 3D reconstruction. Therefore, we propose a method that evaluates the reliability of feature points using epipolar constraint equations and optical flow for matching points of images obtained from a monocular camera, and classifies whether they are correct matching points by machine learning. The specific methods are 1) image matching, 2) calculating epipolar constraint formulas and distances between matching points, and 3) classifying whether matching points are correct by machine learning. 4) Perform 3D reconstruction using only correct matching points. In order to demonstrate the effectiveness of this method, we conducted experiments using simulation images with known three-dimensional structures. For machine classification, K-means clustering method and nonlinear SVM (Support Vector Machine) were used for comparison. We also conducted a similar experiment with a real object. These results suggest that the method can perform correct 3D reconstruction even in the gastrointestinal tract and contribute to the identification of lesions such as early esophageal cancer.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Fumiya Okuno, Kayo Yoshimoto, Hideya Takahashi, and Goro Hamagishi "Evaluation method of matching points for improving accuracy of position estimation using monocular endoscope", Proc. SPIE 12820, Endoscopic Microscopy XIX, 128200A (12 March 2024); https://doi.org/10.1117/12.3001311

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available