Chapter Tracking of Miniature-Sized Objects in 3D Endoscopic Vision Khanam Z., Raheja J.L. (2018) In: Das S., Chaki N. (eds) Algorithms and Applications. Smart Innovation, Systems and Technologies, vol 88. Springer, Singapore.
The advent of 3D endoscope has revolutionized the field of industrial and medical inspection. It allows visual examination of inaccessible areas like underground pipes and human cavity. Miniature-sized objects like kidney stone and industrial waste products like slags can easily be monitored using 3D endoscope. In this paper, we present a technique to track small objects in 3D endoscopic vision using feature detectors. The proposed methodology uses the input of the operator to segment the target in order to extract reliable and stable features. Grow-cut algorithm is used for interactive segmentation to segment the object in one of the frames and later on, sparse correspondence is performed using SURF feature detectors. SURF feature detection based tracking algorithm is extended to track the object in the stereo endoscopic frames. The evaluation of the proposed technique is done by quantitatively analyzing its performance in two ex vivo environment and subjecting the target to various conditions like deformation, change in illumination, and scale and rotation transformation due to movement of endoscope.