% Chapter 7

\chapter{Testbed Design for Comparative Performance Evaluation} % Main chapter title

\label{Chapter7} % For referencing the chapter elsewhere, use \ref{Chapter1} 

\lhead{Chapter 7. \emph{Testbed Design for Comparative Performance Evaluation}} % This is for the header on each page - perhaps a shortened title

%----------------------------------------------------------------------------------------

One of the major difficulties of implementing any capsule localization algorithms when it comes inside the human body is validation. That's because we have limited control of the capsule after it is swallowed by the patient so we could not verify the performance of the algorithms. Besides, carrying out experiments on the real human beings is extremely costly and restricted by law. Thus, the only way to test the localization algorithms is to build emulation testbed. In this chapter, we talk about how to establish a realistic testbed for performance evaluation.

\section{Visual Component}

In the past decade, miniaturization and cost reduction of semiconductor devices have allowed the design of small, low cost computing and wireless communication devices used as sensors in a variety of popular wireless networking applications and this trend is expected to continue in the next few decades. One of the leading wonders of this wireless networking breakthrough is the emergence of wireless video capsule endoscopy (VCE). The technology was introduced by the Given Imaging, Israel in 2000.and the U.S. Food and Drug Administration (FDA) approved its clinical use in 2001. Examination of the Gastro-Intestinal (GI) tract using VCE is commonly used for a number of diseases such as the inflammatory bowel disease, the ulcerative colitis and the colorectal cancer. VCE provides a unique visualization of bleedings and tumors in the middle parts of the small intestine, where traditional endoscopy and colonoscopy visualization techniques cannot reach. Small intestine is a 5-8 meters long curled path occupying the central part of the GI tract, therefore, the path of movement of the VCE inside the small intestineis very complex. Wireless video capsule endoscopy has been in use for over a decade and it uses radio frequency (RF) signals to transmit approximately fifty five thousands clear pictures of inside the GI tract to the body-mounted sensor array.

Wireless capsule endoscopy begins with the patient swallowing the capsule. The natural peristalsis moves the capsule smoothly and painlessly throughout the GI tract while it is transmitting color images taken by the camera at a rate of two images per second [26]. The VCE images allow the physician to visualize the entire GI tract without scope trauma and air insufflations. Comparing with the traditional gastroscopy and colonoscopy techniques, the VCE procedure is ambulatory allowing the patients to continue with their daily activities throughout the endoscopic examinations. In addition, the traditional techniques can only reach the first few or last several feet of the small intestine, while VCE provides images of the entire GI tract.

However, physician has no clue on the exact location of the capsule inside the GI tract to associate it with the pictures showing abnormalities such as bleeding or tumors. It is desirable to use the same RF signal for localization of the VCE as it passes through the human GI tract.

%----------------------------------------------------------------------------------------

\section{RF Component}

In recent years, the feasibility of several technologies for localization of the VCE has been explored. These technologies can be divided into those using magnetic field or inertial systems, using image processing techniques and techniques using RF signals. In magnetic sensing based techniques, a magnet is inserted into the VCE and the VCE is located by measuring the magnetic field. This technique increases the weight and size of the VCE and the magnetic field of the VCE used for localization will be interfered by the external magnetic fields used for other applications such as the Magnetic Resonance Imaging (MRI) systems. One can also insert radiation opaque material into the VCE and trace the location of the VCE using X-ray or Computed Tomography (CT) scan. Continuous imaging using X-ray or CT scan is very expensive and it bears the health risks for the patient. Using the RF signal used for image transmissions for the VCE to also locate the capsule offers itself as a natural and low cost solution that does not add to the capsule complexity and payload. Therefore, it has been chosen for use with the smartpill capsule in USA and the M2A capsule in Israel. These companies use the RSS of the waveform for the purpose of localization of the VCE. A more accurate metric for localization is the TOA or the time of flight of the signal.

For RF based localization, a widely known benefit of TOA based techniques is their high accuracy compared to RSS based techniques. The TOA based technique relies on measurements of travel time of signals between the known reference nodes and unknown terminal nodes. Ranging information is calculated by multiplying the propagation velocity of RF signal and the measured TOA value. The testbed developed in [40] can be used to examine the performance of TOA localization in indoor areas under the influence of multipath scenario. On the other hand, the human body is formed of various organs with complex structures. Each organ has unique characteristics of conductivity and relative permittivity. Since propagation velocity inside human body is expressed as a function of the relative permittivity, medical implanted devices placed in different positions cause different propagation velocities due to the RF signal traveling through various tissues or organs. These variations in the speed are dominant source of error for TOA-based RF localization inside the human body. However, there is nothing available in the literature to compare the performances of the performances of the two approaches in particular in a realistic 3D scenario.

%----------------------------------------------------------------------------------------

\section{Performance Evaluation for Body-SLAM}

Wireless capsule endoscopy (WCE) is progressively emerging as a popular non-invasive imaging tool for gastrointestinal (GI) tract diagnosis. Compared with the traditional colonscope or enterscope, WCE has the capability of examining the entire small intestine, which other endoscopic instruments can not reach. However, since the length of small intestine is too long (varies from 5m to 9m and it is twisted inside the abdominal cavity with extremely indistinguishable distribution, the localization of the capsule inside small intestine becomes very challenging, which prevents physicians from administering immediate therapeutic operations after an abnormality is found by the video source. Thus, having a precise and reliable localization system for the capsule inside the small intestine would greatly enhance the benefits of WCE. 

During the past few years, many attempts have been made to develop accurate and reliable localization systems for the WCE. A commonly used localization infrastructure, which has been chosen for commercial use for M2A capsule designed by Given Imaging, is to attach many calibrated external antennas to the anterior abdominal wall of the human body to detect the RF signal emitted by the wireless capsule . By interpreting the power of the received signal into distance between the capsule and body mounted sensor array, position of the capsule can be estimated by pattern matching algorithms such as least square algorithm and maximum likelihood algorithm . However, due to the non-homogeneity and severe attenuation of body tissues, features of the received signal are sometimes poorly correlated with the distance. Therefore, this RF localization system often end up providing discontinuous and scattered estimations with unacceptable amount of error.

One way to enhance the performance of RF localization is to combine the motion information of the capsule by employing a data fusion algorithm such like Kalman filter or particle filter. In the localization literature, there has been a trend to extract motion parameters from  image sequence to improve the accuracy of RF localization. This class of algorithms is known as video based simultaneous localization and mapping (SLAM) algorithms . In the WCE application, since the endoscopic capsule continually takes pictures with very short time interval (two frames / sec), it is possible to reconstruct motion information of the capsule from video stream . In this paper, we present a hybrid localization technique that is able to extract speed and moving direction of the endoscopic capsule from endoscopic image frames to aid the RF localization. The major contribution of this paper is that we explored the potential of using images as another source to track the position of WCE and we established a virtual platform to validate our algorithm.

%----------------------------------------------------------------------------------------