Camera-Based Surgical Navigation System: Evaluation of Classification and Object Detection CNN Models for X-markers Detection

Gieseler, Oliver; Gamal, Omar; Kumar, Shashank; Roth, Hubert

doi:10.1109/icmre51691.2021.9384815

Cited by 1 publication

(1 citation statement)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the advent of deep learning, several related methods have arisen in the domain of X corner detection, such as Gieseler [40], which treated X-corner detection as an object detection task and employed CNN for detection. Although deep learning models with a large number of parameters have shown the potential for performance improvement, the real-time requirements of 3D tracking tasks cannot be met without highperformance hardware resources.…”

Section: Introductionmentioning

confidence: 99%

Robust 3D Object Tracking with X-Triplet Markers: A Machine Learning-Based Approach

Shen

Cai

Deng

et al. 2023

Preprint

View full text Add to dashboard Cite

In this paper, we propose an efficient and robust X-Triplet detection method based on the support vector machine (SVM) and an adjacent matrix for locating and tracking objects through stereo vision with minimal feature points. The X-Triplet, denoted as Tri-X, is a composite marker consisting of three sequential X-corners. The definition and types of Tri-X markers are initially presented. Further, a fast and robust X-corner detector based on the block search strategy and SVM is proposed for extracting X-corner candidates along with their sub-pixel locations and orientations. The X-corner adjacent matrix (XAM) is then constructed using the orientation angle error to describe the possibility that any X-corner pair can form a valid edge vector. The Tri-X candidates are extracted efficiently from the XAM. Once the Tri-X markers are detected in binocular images, their 6D pose information can be recovered through stereo matching and triangulation technique. When multiple targets are involved simultaneously, different Tri-X markers can be utilized to identify different objects. Experimental results demonstrate the superiority of our method in terms of both accuracy and efficiency for X-corner and Tri-X marker detection, surpassing state-of-the-art approaches. Notably, our method achieves exceptional localization precision with a remarkable position error below 0.1mm and an orientation error below 1 degree. Additionally, our method exhibits great potential for utilization in user-defined specific tracking tasks, offering flexibility and adaptability to various tracking requirements.

show abstract

Section: Introductionmentioning

confidence: 99%