Search

Purpose: To propose a novel approach for automated measurements of margin reflex distance (MRD) using digital image analysis and to evaluate the agreement between automated and manual measurements of MRD. Materials and Methods: This observational study included 132 eyes of 66 volunteers referred to an oculoplastic clinic for blepharoptosis. Preoperative facial photographs of participants were taken. MRD1 and MRD2 were measured manually by a senior surgeon and automatically by our image-based algorithm. Correlation analyses and Bland-Altman analyses were performed to evaluate the agreement between the two measurements. Results: A strong positive correlation was observed between automated and manual MRD1 measurements, with Spearman's r = 0.968 (95% confidence interval [CI] = 0.949-0.976; P < .001) and intraclass correlation coefficient (ICC) = 0.966 (95% CI = 0.953-0.976; P < .001). The bias between automated and manual MRD1 measurements was 0.02 mm (95% CI = -0.06-0.10 mm), with 95% limits of agreement (LoA) = -0.88 to 0.92 mm. Agreement also existed in MRD2 measurements, with Spearman's r = 0.803 (95% CI = 0.716-0.865; P < .001), ICC = 0.833 (95% CI = 0.772-0.879; P < .001), and bias = 0.34 mm (95% CI = 0.26-0.42 mm), 95% LoA = -0.54 to 1.22 mm. Conclusions: Automated eyelid measurements in blepharoptosis using the image-based approach compare favorably with clinical manual measurements. This novel approach allows an objective assessment of ptosis with high accuracy.

Background: Patients with non-calcified hamartoma were more susceptible to suffer from surgery or needle biopsy for the tough discrimination from lung adenocarcinoma. Radiomics have the ability to quantify the lesion features and potentially improve disease diagnosis. Thus, this study aimed to discriminate non-calcified hamartoma from adenocarcinoma by employing imaging quantification and machine learning. Methods: Forty-two patients with non-calcified hamartoma and 49 patients with adenocarcinoma were retrospectively collected from our institute, that were confirmed by pathology. Manual lesion segmentation, feature quantification (e.g., texture features) and artificial neural network were performed consecutively. Independent t-test was used to conduct the inter-group comparisons of those imaging features. Receiver operating characteristic curve was performed to investigate the discriminating efficacy. Results: Significantly higher contrast, cluster prominence, cluster shade, dissimilarity, energy and entropy in non-calcified hamartoma were observed compared with lung adenocarcinoma. Texture-grey-level co-occurrence matrix showed a well discrimination between non-calcified hamartoma and adenocarcinoma as the detection sensitivity, specificity, accuracy and the area under the curve were 87.22% ± 9.07%, 82.64% ± 8.07%, 85.11% ± 5.40% and 0.942, respectively. Conclusion: Quantifying imaging features is a potentially useful tool for clinical diagnosis. This study demonstrated that non-calcified hamartoma has a heterogeneous distribution of attenuations probably resulting from its complex organizations. And based on this property, imaging quantification could improve discrimination of non-calcified hamartoma from adenocarcinoma.

We have developed a new computed tomography (CT)-based 3-D volume difference reconstruction (VDR) method for building a model of soft tissue differences using facial self-referenced prototyping computation, and a computer-aided design (CAD) workflow for fabricating custom ocular prostheses (COPs). In facial self-referenced prototyping computation, an ocular prosthesis is constructed from a combination of a top surface, shape contour, and bottom surface, with key measurements being made with respect to the patient's healthy eye. The top surface is derived from the outer region of the healthy eyeball, using Hough transformation. In the meantime, the shape contour and bottom surface are computed from facial edges fitted to Fourier curves. Once these self-referenced parameters have been obtained, finite element analysis is used to generate a primitive prosthesis model. Before being converted into a format suitable for rapid prototyping techniques, the model is refined using surface equations. Oculists choose the final optimized model for printing. Geometric errors, compared with a manually-fabricated prosthesis which has gone through multiple rounds of time-consuming manual adjustment, are as low as 0.46%, and the fabrication time and human error have been greatly reduced. The fill density of the COP is 60%, the wall thickness is 1.6 mm and the weight is 1.59 g. The reconstructed contours and symmetry of the COP are satisfactory with doctors' evaluation and the patient is able to have a normal appearance. The CT-based VDR method we first proposed here can accurately build a model of the soft tissue difference. The full CAD workflow has improved prosthesis design and manufacturing efficiency. This flow can also be applied to other soft issue reconstruction.

Stomach bleeding is a kind of gastrointestinal disease which can be diagnosed noninvasively by wireless capsule endoscopy (WCE). However, it requires much time for physicians to scan large amount of WCE images. Alternatively, computer-assisted bleeding localization systems are developed where color, edge, and intensity features are defined to distinguish lesions from normal tissues. This paper proposes a saliency-based localization system where three saliency maps are computed: phase congruency-based edge saliency map derived from Log-Gabor filter bands, intensity histogram-guided intensity saliency map, and red proportion-based saliency map. Fusing the three maps together, the proposed system can detect bleeding regions by thresholding the fused saliency map. Results demonstrate the accuracy of 98.97% for our system to mark bleeding regions.

Objective: Many common eye diseases and cardiovascular diseases can be diagnosed through retinal imaging. However, due to uneven illumination, image blurring, and low contrast, retinal images with poor quality are not useful for diagnosis, especially in automated image analyzing systems. Here, we propose a new image enhancement method to improve color retinal image luminosity and contrast. Methods: A luminance gain matrix, which is obtained by gamma correction of the value channel in the HSV (hue, saturation, and value) color space, is used to enhance the R, G, and B (red, green and blue) channels, respectively. Contrast is then enhanced in the luminosity channel of L*a*b* color space by CLAHE (contrast-limited adaptive histogram equalization). Image enhancement by the proposed method is compared to other methods by evaluating quality scores of the enhanced images. Results: The performance of the method is mainly validated on a dataset of 961 poor-quality retinal images. Quality assessment (range 0–1) of image enhancement of this poor dataset indicated that our method improved color retinal image quality from an average of 0.0404 (standard deviation 0.0291) up to an average of 0.4565 (standard deviation 0.1000). Conclusion: The proposed method is shown to achieve superior image enhancement compared to contrast enhancement in other color spaces or by other related methods, while simultaneously preserving image naturalness. Significance: This method of color retinal image enhancement may be employed to assist ophthalmologists in more efficient screening of retinal diseases and in development of improved automated image analysis for clinical diagnosis.

Purpose Retinal imaging is an important and effective tool for detecting retinal diseases. However, degraded images caused by the aberrations of the eye can disguise lesions, so that a diseased eye can be mistakenly diagnosed as normal. In this work, we propose a new image enhancement method to improve the quality of degraded images. Methods A new method is used to enhance degraded-quality fundus images. In this method, the image is converted from the input RGB colour space to LAB colour space and then each normalized component is enhanced using contrast-limited adaptive histogram equalization. Human visual system (HVS)-based fundus image quality assessment, combined with diagnosis by experts, is used to evaluate the enhancement. Results The study included 191 degraded-quality fundus photographs of 143 subjects with optic media opacity. Objective quality assessment of image enhancement (range: 0–1) indicated that our method improved colour retinal image quality from an average of 0.0773 (variance 0.0801) to an average of 0.3973 (variance 0.0756). Following enhancement, area under curves (AUC) were 0.996 for the glaucoma classifier, 0.989 for the diabetic retinopathy (DR) classifier, 0.975 for the age-related macular degeneration (AMD) classifier and 0.979 for the other retinal diseases classifier. Conclusion The relatively simple method for enhancing degraded-quality fundus images achieves superior image enhancement, as demonstrated in a qualitative HVS-based image quality assessment. This retinal image enhancement may, therefore, be employed to assist ophthalmologists in more efficient screening of retinal diseases and the development of computer-aided diagnosis.

Background Asymmetric medullary veins (AMV) are frequently observed in stroke patients and single-echo susceptibility weighted imaging (SWIs) is the main technique in detecting AMV. Our study aimed to investigate which echo time (TE) on single-echo susceptibility is the optimal echo for visualizing AMV and to compare the ability in detecting AMV in stroke patients between SWIs and multi-echo susceptibility weighted imaging (SWIc). Materials and methods Twenty patients with middle cerebral artery stroke were included. SWI was acquired by using a multi-echo gradient-echo sequence with six echoes ranging from 5 ms to 35.240 ms. Three different echoes of SWIs including SWIs1 (TE = 23.144 ms), SWIs2 (TE = 29.192 ms) and SWIs3 (TE = 35.240 ms) were reconstructed. SWIc was averaged using the three echoes of SWIs. Image quality and venous contrast of medullary veins were compared between SWIs and SWIc using peak signal-to-noise ratio (PSNR), mean opinion score (MOS), contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR). The presence of AMV was evaluated in each SWIs (1-3) and SWIc. Results SWIs2 had the highest PSNR, MOS and CNR and SWIs1 had the highest SNR among three different echoes of SWIs. No significant difference was found in SNR between SWIs1 and SWIs2. PSNR, MOS and CNR in SWIc were significantly increased by 27.9%, 28.2% and 17.2% compared with SWIs2 and SNR in SWIc was significantly increased by 32.4% compared with SWIs1. 55% of patients with AMV were detected in SWIs2, SWIs3 and SWIc, while 50% AMV were found in SWIs1. Conclusions SWIs using TE around 29ms was optimal in visualizing AMV. SWIc could improve image quality and venous contrast, but was equal to SWIs using a relative long TE in evaluating AMV. These results provide the technique basis for further research of AMV in stroke.

Telemedicine and the medical "big data" era in ophthalmology highlight the use of non-mydriatic ocular fundus photography, which has given rise to indispensable applications of portable fundus cameras. However, in the case of portable fundus photography, non-mydriatic image quality is more vulnerable to distortions, such as uneven illumination, color distortion, blur, and low contrast. Such distortions are called generic quality distortions. This paper proposes an algorithm capable of selecting images of fair generic quality that would be especially useful to assist inexperienced individuals in collecting meaningful and interpretable data with consistency. The algorithm is based on three characteristics of the human visual system--multi-channel sensation, just noticeable blur, and the contrast sensitivity function to detect illumination and color distortion, blur, and low contrast distortion, respectively. A total of 536 retinal images, 280 from proprietary databases and 256 from public databases, were graded independently by one senior and two junior ophthalmologists, such that three partial measures of quality and generic overall quality were classified into two categories. Binary classification was implemented by the support vector machine and the decision tree, and receiver operating characteristic (ROC) curves were obtained and plotted to analyze the performance of the proposed algorithm. The experimental results revealed that the generic overall quality classification achieved a sensitivity of 87.45% at a specificity of 91.66%, with an area under the ROC curve of 0.9452, indicating the value of applying the algorithm, which is based on the human vision system, to assess the image quality of non-mydriatic photography, especially for low-cost ophthalmological telemedicine applications.

A novel full-reference image quality assessment metric employing statistical local correlation as an indicator for quality quantification is presented. The local correlation is extracted in a wavelet domain and is pooled into an objective quality score. Simulation and testing results demonstrate that the proposed method not only outperforms the other methods in terms of high accuracy of image quality prediction, but also is consistent with the subjective evaluations.

A generalized relative quality (RQ) assessment scheme is proposed here based on the Bayesian inference theory, which is reasonable to make use of full reference (FR) algorithms when the evaluation of the quality of homogeneous medical images is required. Each FR algorithm is taken as a kernel to represent the level of quality. Although, various kernels generate different order of magnitude, a normalization process can rationalize the quality index within 0 and 1, where 1 represent the highest quality and 0 represents the lowest quality. To validate the performance of the proposed scheme, a series of reconstructed susceptibility weighted imaging images are collected, where each image has its subjective scale. Both experimental results and a ROC analysis show that the RQ obtained from the proposed scheme is consistent with subjective evaluation.

Laser ultrasonic technology has been paid more and more attention in the NDT field due to its outstanding ability of non-contact measurements. As a good vibration measurement technology, CFPI (confocal Fabry-Perot interferometer) method is sensitive to the laser frequency shift caused by the ultrasonic wave propagation. In this paper, a laser ultrasonic receiver system based on the CFPI method is established, and the system parameters are analyzed in order to obtain an optimal resolution. Also some initial experiment results are presented to certify this system ability.

A no-reference objective metric for image quality assessment by integrating the topographic independent components analysis into feature extraction is presented. By taking the topographic relationship among the initially independent features into consideration, it extracts the features of more sparsity or independency which is essentially related to inherent quality. Evaluation results demonstrate that the proposed metric is able to predict the image quality accurately across various distortion types.