ASIFT: A New Framework for Fully Affine Invariant Image Comparison
SIAM Journal on Imaging Sciences
Hi-index | 0.00 |
Designing a precise and reliable localization system for wireless capsule endoscopy (WCE) has always been a challenging problem due to the complicated in-body environment and uncontrollable movement of body tissue. Knowing the motion information of the capsule would greatly enhance the localization accuracy. However, design and validate any motion tracking algorithm inside small intestine faces a lot of difficulties since any experimentation on the human being is extremely costly and restricted by law. Having a virtual environment that looks and functions exactly like small intestine would facilitate the process of verifying the performance of existing algorithms without going into the real human body. In this paper, we establish a virtual testbed that emulates the contraction of intestinal lumen and the transition of endoscopic capsule. Under this emulation environment, a velocity estimation algorithm based on a feature detection algorithm (ASIFT) and on a velocity estimation algorithm (MDR) is designed and implemented and its performance has been evaluated. Experimental results showed that our proposed emulation environment is able to provide reliable platform for motion detection validation.