DEPTH ESTIMATION METHOD AND DEVICE FOR ENDOSCOPIC IMAGE

• Main Authors: YOON, Ju Hong, PARK, Min Gyu
• Format: Patent
• Language: English; French; Korean
• Published: 10.06.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DIAGNOSIS; HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATIONTECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING ORPROCESSING OF MEDICAL OR HEALTHCARE DATA; HUMAN NECESSITIES; HYGIENE; IDENTIFICATION; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; MEDICAL OR VETERINARY SCIENCE; PHYSICS; SURGERY

Provided are a depth estimation method and device for an endoscopic image. The depth estimation method, according to one embodiment of the present invention, comprises generating a monocular image, and receiving the generated monocular image and predicting a direct attenuation model (DAM)-based disparity map by using a deep learning network for estimating DAM-based disparity maps, wherein DAM is a model for estimating a disparity map from a monocular image on the basis that light of lighting used for generating the monocular image tends to be attenuated. L'invention concerne un procédé et un dispositif d'estimation de profondeur pour une image endoscopique. Le procédé d'estimation de profondeur, selon un mode de réalisation de la présente invention, consiste à produire une image monoculaire, puis à recevoir l'image monoculaire générée et à prédire un mappage de disparité à base de modèle d'atténuation directe (DAM) à l'aide d'un réseau d'apprentissage profond pour estimer des mappages de disparité à base de DAM. Un DAM est un modèle d'estimation d'un mappage de disparité à partir d'une image monoculaire sur la base de la tendance à l'atténuation de la lumière d'éclairage utilisée pour générer l'image monoculaire. 내시경 영상에 대한 깊이 추정 방법 및 장치가 제공된다. 본 발명의 실시예에 따른 깊이 추정 방법은, 단안 영상을 생성하고, 생성된 단안 영상을 입력받아 DAM(Direct Attenuation Model) 기반 디스패리티 맵을 추정하는 딥러닝 네트워크를 이용하여 DAM 기반 디스패리티 맵을 예측하며, DAM은 단안 영상을 생성하는데 이용되는 조명의 빛이 감쇄되는 경향성을 기초로 단안 영상으로부터 디스패리티 맵을 추정하는 모델이다.