Errors In Image Analysis Research Articles

Voxel-based superimposition of Cone Beam CT scans for orthodontic and craniofacial follow-up: Overview and clinical implementation

The increasing use of three-dimensional (3D) imaging in orthodontics has led to the development of 3D superimposition techniques. These techniques use stable anatomic structures as references in order to compare Cone Beam CT (CBCT) scans of the same subject at different time-points. Three methods have been described in the literature: landmark-based, surface-based and voxel-based 3D superimpositions. This article focuses on the voxel-based approach, which is the most described and the only one that can be fully automatized. The aim of this paper is to offer clinicians a practical tutorial on craniofacial voxel-based 3D superimposition. We provide an updated overview of the available implementation methods, describing their methodology, validations, main steps, advantages and drawbacks. The historical open-source method is the most widespread for research purposes, but takes around three hours to achieve for an experienced operator. Several commercially-available software perform superimpositions in a few minutes. We used two of the available methods to conduct the superimposition process with three representative clinical cases in order to illustrate the different types of results that can be obtained. Commercially-available software provide user-friendly and fully automatized superimposition methods, allowing clinicians to perform it easily and helping to reduce human error in image analysis. Still, quantitative evaluation of the results remains the main challenge of this technique.

Accounting for optical errors in microtensiometry

HypothesisDrop shape analysis (DSA) techniques measure interfacial tension subject to error in image analysis and the optical system. While considerable efforts have been made to minimize image analysis errors, very little work has treated optical errors. There are two main sources of error when considering the optical system: the angle of misalignment and the choice of focal plane. Due to the convoluted nature of these sources, small angles of misalignment can lead to large errors in measured curvature. We demonstrate using microtensiometry the contributions of these sources to measured errors in radius, and, more importantly, deconvolute the effects of misalignment and focal plane. Our findings are expected to have broad implications on all optical techniques measuring interfacial curvature. ExperimentsA geometric model is developed to analytically determine the contributions of misalignment angle and choice of focal plane on measurement error for spherical cap interfaces. This work utilizes a microtensiometer to validate the geometric model and to quantify the effect of both sources of error. For the case of a microtensiometer, an empirical calibration is demonstrated that corrects for optical errors and drastically simplifies implementation. FindingsThe combination of geometric modeling and experimental results reveal a convoluted relationship between the true and measured interfacial radius as a function of the misalignment angle and choice of focal plane. The validated geometric model produces a full operating window that is strongly dependent on the capillary radius and spherical cap height. In all cases, the contribution of optical errors is minimized when the height of the spherical cap is equivalent to the capillary radius, i.e. a hemispherical interface. The understanding of these errors allow for correct measure of interfacial curvature and interfacial tension regardless of experimental setup. For the case of microtensiometry, this greatly decreases the time for experimental setup and increases experiential accuracy. In a broad sense, this work outlines the importance of optical errors in all DSA techniques. More specifically, these results have important implications for all microscale and microfluidic measurements of interface curvature.

Binarization Method Based on a Local Maximum-Ratio Threshold

For a vision-based analysis system, noise of the smear, the strain, or the badly illumination often causes serious errors in image analysis due to misclassification of object pixels and background pixels. An effective binarization method is thus required to improve the accuracy of segmentation and this paper proposes a novel algorithm for determining the threshold of each pixels. The proposed approach does not require statistical computation to determine the local threshold. The thresholding value can be easily determined by using the ratio of maximum intensity in a square region around each pixel. Experimental studies are conducted to evaluate the performance of the proposed method and the obtained results verify its segmentation abilities of hand-written text and music symbol under the badly illumination condition.

Image-Based Traffic Monitoring With Shadow Suppression

For a vision-based traffic monitoring and enforcement system, shadows of moving objects often cause serious errors in image analysis due to misclassification of shadows and moving vehicles. An effective shadow suppression method is thus required to improve the accuracy of image analysis and this paper proposes a novel color-space ratio model for detecting shadow pixels in traffic imagery. The proposed approach does not require many image sequences for constructing the model. Instead the model can be easily built up using a shadow region in a single image frame. To increase the accuracy of shadow detection, we design two types of spatial analysis to verify actual shadow pixels. Comparative results show that the proposed method works better than several well-known methods. The proposed methods have been applied to an image-based traffic monitoring system for detecting shadow pixels in traffic imagery. The experimental results not only validate the feasibility of the proposed algorithm but also successfully estimate traffic parameters such as traffic flows, traffic densities, vehicle turn ratios and vehicle speeds, all with satisfactory accuracy

Peripapillary Nerve Fiber Layer Thickness Measurement Reproducibility Using Scanning Laser Polarimetry

The authors assess the reproducibility of retinal nerve fiber layer (RNFL) thickness measurements using scanning laser polarimetry. The right eyes of five healthy individuals were scanned by two trained operators (S.T.H., H.I.) using the Nerve Fiber Analyzer II (Laser Diagnostic Technologies, Inc., San Diego, CA, U.S.A.). Each participant was scanned six consecutive times per session for five sessions carried out over separate days within a two-week period. Intra- and interoperator measurement reproducibility was assessed. Intraoperator reproducibility was high. Mean coefficients of variation for total RNFL thickness measurements were 4.48 +/- 1.76% and 4.92 +/- 2.32% for operators 1 and 2, respectively. Interoperator reproducibility was high (p = 0.20-0.93) if a single ellipse was applied to the images of both examiners. However, measurement reproducibility worsened if each operator created his own measurement ellipse (p < or = 0.05 for 3 out of 5 participants). Interoperator error in image analysis can be minimized by using a single ellipse for the baseline and all subsequent scans even if subsequent scans are acquired by different operators.