International journal of computer assisted radiology and surgery · 2026
Enhanced distance perception in automated robotic endoscopy through scale-aware monocular depth estimation.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Wei Ruofeng, Chen Kai, Ma Yiyao, Li Bin, Liu Yunhui, Dou Qi
Аннотация
PURPOSE: The increasing demand for distance perception in surgical robotic systems is crucial for ensuring safe and effective operations in minimally invasive procedures. Accurate distance awareness allows surgical robots to autonomously navigate complex anatomical structures, particularly in endoscopic environments with limited visual feedback. Existing distance estimation methods often rely on expensive sensors that can pose safety risks and are ineffective under these conditions. To address this challenge, this work develops a monocular image-based framework leveraging geometric modeling and long-term tracking technologies to enable accurate distance measurements from a monocular camera, enhancing current endoscopic practices.
METHODS: The proposed approach introduces an endoscopic image-based pipeline designed to estimate high-quality, scale-aware depth from monocular endoscopic scenes. This pipeline utilizes surgical instruments with cylindrical shafts of known radius as geometric constraints to recover absolute scale. An automatic module is designed for continuous tracking of specific tissue regions and instruments within the scene, allowing for precise distance calculations between various objects.
RESULTS: Compared to state-of-the-art baselines, the depth model reduces mean absolute error from 8.56 mm to 6.35 mm and improves accuracy from 76.9% to 88.0%, while achieving near-perfect absolute scale estimation. The overall framework attains average distance errors between 4.7 and 6.4 mm across multiple measurement types, demonstrating reliable and efficient performance in complex endoscopic scenarios.
CONCLUSION: We present a novel image-based distance perception framework aimed at enhancing surgical robot automation. The framework facilitates efficient measurements of various objects within endoscopic views. Through multiple distance-augmented endoscopic videos, we demonstrate the potential of our work to benefit robotic surgery tasks, including automatic navigation.
METHODS: The proposed approach introduces an endoscopic image-based pipeline designed to estimate high-quality, scale-aware depth from monocular endoscopic scenes. This pipeline utilizes surgical instruments with cylindrical shafts of known radius as geometric constraints to recover absolute scale. An automatic module is designed for continuous tracking of specific tissue regions and instruments within the scene, allowing for precise distance calculations between various objects.
RESULTS: Compared to state-of-the-art baselines, the depth model reduces mean absolute error from 8.56 mm to 6.35 mm and improves accuracy from 76.9% to 88.0%, while achieving near-perfect absolute scale estimation. The overall framework attains average distance errors between 4.7 and 6.4 mm across multiple measurement types, demonstrating reliable and efficient performance in complex endoscopic scenarios.
CONCLUSION: We present a novel image-based distance perception framework aimed at enhancing surgical robot automation. The framework facilitates efficient measurements of various objects within endoscopic views. Through multiple distance-augmented endoscopic videos, we demonstrate the potential of our work to benefit robotic surgery tasks, including automatic navigation.