ABSTRACT Mitral valve disease is a commonly occurring pathology and its incidence is expected to rise in the coming decades. This has led to an increase in the number and complexity of computational models developed to characterize the mitral valve behaviour, resulting in the need for automated segmentation methods. In this study, we evaluated the suitability of particle systems for segmenting the mitral valve leaflets from 3D TOE images, for both systolic and diastolic phases. With a mean error of 0.77 ± 0.4 mm, our approach results in accurate medial surface representations of the mitral valve. Additionally, we investigated the correlation between particle strength and leaflet thickness. Although a thorough analysis is still required to fully establish this relation, our preliminary results, with a mean RMS error of 0.99 ± 0.27 mm, show that leaflet thickness can be linked to particle strength. We anticipate that particle systems can be of high relevance for the generation of realistic patient-specific computational models, contributing to the development of adequate treatment strategies for mitral valve patients.

ABSTRACT Membranous nephropathy (MN) is among the most common glomerular diseases that cause nephrotic syndrome in adults. To aid pathologists on performing the MN classification task, we proposed here a pipeline consisted of two steps. Firstly, we assessed four deep-learning-based architectures, namely, ResNet-18, MobileNet, DenseNet, and Wide-ResNet. To achieve more reliable predictions, we adopted and extensively evaluated a Monte-Carlo dropout approach for uncertainty estimation. Using a 10-fold cross-validation setup, all models achieved average F1-scores above 92%, where the highest average value of 93.2% was obtained by using Wide-ResNet. Regarding uncertainty estimation with Wide-ResNet, high uncertainty scores were more associated with erroneous predictions, demonstrating that our approach can assist pathologists in interpreting the predictions with high reliability. We show that uncertainty-based thresholds for decision referral can greatly improve classification performance, increasing the accuracy up to 96%. Finally, we investigated how the uncertainty scores relate to complexity scores defined by pathologists.

ABSTRACT Hand telerehabilitation currently has limitations for accurate and remote assessment of range of motion (ROM) in small finger joints. ‘DIGITS’ application utilises the front smartphone camera to measure finger ROM in a reliable and rapid assessment protocol. Our initial beta-phase testing examined the consistency of our software measurements to in-person goniometry. 6 to 9 degrees of difference existed between the smartphone application recorded data versus the in-person measurements. This range is within acceptable 7 to 9 degree tolerance for interrater goniometry measurements. The effect of environmental factors such as hand distance, lightings and hand orientation was evaluated. The intraclass correlation coefficient (ICC) was 0.98–0.98 for finger extension and 0.78–0.81 for finger flexion recorded across different environmental conditions. Overall, ‘DIGITS’ provides an augmented reality tool to generate reliable finger ROM tracking for hand telerehabilitation. ‘DIGITS’ has the potential to transform hand therapy and improve our knowledge of hand pathology recovery.

ABSTRACT Gaussian filtering is a successful computer operation vision to reduce noise and calculate the gradient intensity change of an image. However, it’s well known that in scale space context, the Gaussian kernel has some drawbacks, loss of information caused by the unavoidable Gaussian truncation and the prohibitive processing time due to the mask size. To give a solution to both problems, a new kernel family with compact support and its separable version were presented in the literature. The theoretical study of these kernels shows that the new family kernel is parameterised by a scale parameter and generated in such a way that fine scale structures are successively suppressed when the scale parameter is increased. Moreover, the scale parameter is increased, the image is blurred and details and border are removed. All these disadvantages are related to the static nature of these kernels. In this paper, we propose a smart kernel based on deep neural networks (dnn) to create a dynamic kernel with compact support called DSKCS. The parameter involved in the filtering process is calculated in real time and supervised by deep neural networks. The filter is continuously variable to detect, clean and avoid noisy areas of the image. Extensive experiments show that the proposed kernel can improve the classic kernel and presents a solution for its limitations related to its static nature. Furthermore, different metrics calculated illustrate our approach efficiency. As stated in the filtering performance, which reveals the highest PSNR and NCC with the metrics results (PSNR = 32.18, NCC = 0.95). Also, we recorded more than 0.89 for area under curves of the classification results using DBN-DNN technique.

ABSTRACT Acne is a common skin lesion, which has an important impact on people’s lives. Skin analysis is one of the important steps before starting treatment. The purpose of this paper is to build a robot, which detects and burns the skin lesions by the Plexr Plus machine, automatically. Detection of skin lesions localisation is the first and the most important step in treating them. In this study, within the training phase, we used the Visio Face machine as a training image database because its images are very similar to those of Kinect. In the following phase, feature extraction step, the images’ colour properties are used to identify the lesions which are done in RGB colour space. In addition, we used the Genetic Algorithm for selecting the best features. In the test phase, skin lesions are identified in the RGB image of the Kinect camera. Then, the three-dimensional position of lesions on the face is identified by using a depth sensor and the RGB image of the Kinect camera. After that three-dimensional data is sent into a robot designed for treating skin lesions. Finally, we examine the previous methods on the images of the Kinect camera. The results presented 98% accuracy, 97% specificity, 91% precision and 18% sensitivity on average.

ABSTRACT Parkinson’s disease (PD) is an incurable long-term neurodegenerative disorder that mainly influence the motor system and eventually results in significant morbidity. The use of computational tools for classification and prediction of PD is desirable for detecting the symptoms at the onset. In this paper, we developed a novel machine learning technique based on incremental support vector machine and modified Frank-Wolfe method (SVM-MFW) for classification and prediction of PD onset. The proposed model works on heterogeneous data gathered from the Parkinson’s progression markers initiative (PPMI) data repository. The work when compared to earlier state-of-the-art techniques outplays them in terms of both classification and prediction. It is shown that the model achieves an equivalent accuracy of 98.3% and a low cross-entropy value in less computational time.

ABSTRACT This paper reveals an automatic approach for macular diagnostic by the use of Optical Coherence Tomography (OCT) in the evaluation process. At preliminary behaviour levels, the evaluation of macular zone in OCT scan is characterised by an error prone task, related to the experience and the attention of ophthalmologists. Thus, different techniques of OCT-image analysis help the option of obtaining a consistent and independent diagnosis to identify macular degeneration behaviour. In this work, we report an automated approach based on a combined filtering and classification strategy. The presented method is validated on a real integrated diabetic oedema macular (DME) and age related to macular degeneration (AMD). The experimental results illustrate the high accuracy of the results of the proposed method compared to the ground truth. Furthermore, a comparative study with existing techniques is presented in order to demonstrate the efficiency and the superiority of the proposed technique.

ABSTRACT Accurate segmentation of infrared thermal images is a challenging issue in the breast cancer detection. This paper presents and evaluates novel segmentation method for breast cancer detection using infrared thermal images. The developed segmentation method uses developed segmentation algorithm that is hybrid of Gaussian Mean Shift (GMS) and roulette wheel selection approach. In the first stage of the developed segmentation method, the redundant portions of the infrared thermal image are removed and then, the infrared thermal image is divided into five and six clusters by applying the developed segmentation algorithm. The segmented infrared thermal image is created by multiplying the selected best image of each cluster. This image is used to diagnosis the normal breast from abnormal breast. The proposed method is developed using MATLAB 2017 and infrared thermal images of 64 patients. The results show that the average Dice similarity coefficient, Jaccard index and Hausdorff distance in the proposed segmentation method are 91.81%, 84.86%, and 4.87, respectively. The comparison results demonstrate that using the proposed segmentation method can improve the performance of the Computer-Aided-Detection (CAD) system compared to the CAD systems that use the Mean Shift (MS) and Fuzzy C-Means (FCM) segmentation algorithms.

ABSTRACT The aim of this study was to develop and verify a personalised approach for simulating the aerodynamic flow patterns using patient-specific data of a severe form of nasal obstruction during inspiration. The reconstruction of the 3D geometry of the nasal cavity based on the computer images of a patient diagnosed with the unilateral nasal valve collapse and S-shaped septum deviation was used to perform the Computational Fluid Dynamics (CFD) simulation of airflow. A comparative study of the airflow with standard data (SD) and personal input data (PD) of mass flow rate was carried out. In both cases, the air streamlines went straight along the lower and middle nasal passages on the non-deviated septum side. The peak velocity increased in two cavities, mostly on the septum deviation side, while the peak pressure inversely dropped. . With personal data, the peak velocity and pressure values varied greatly in different parts of the nasal cavity, depending on the particular local anatomy. In a patient with unilateral nasal valve collapse and S-shaped septum deviation, the use of personal data does not have an impact on the streamline’s direction, longitudinal velocity, and pressure distributions but it significantly changed the local velocity and pressure in the nasal cavity.

ABSTRACT Safe and efficient surgical training and workflow management play a critical role in clinical competency and ultimately, patient outcomes. Video data in minimally invasive surgery (MIS) have enabled opportunities for vision-based artificial intelligence (AI) systems to improve surgical skills training and assurance through post-operative video analysis and development of real-time computer-assisted interventions (CAI). Despite the availability of mounted cameras for the operating room (OR), similar capabilities are much more complex to develop for recording open surgery procedures, which has resulted in a shortage of exemplar video-based training materials. In this paper, we present a potential solution to record open surgical procedures using head-mounted cameras. Recorded videos were anonymised to remove patient and staff identifiable information using a machine learning algorithm that achieves state-of-the-art results on the OR Face dataset. We then propose a CNN-LSTM-based model to automatically segment videos into different surgical phases, which has never been previously demonstrated in open procedures. The redacted videos, along with the automatically predicted phases, are then available for surgeons and their teams for post-operative review and analysis. To our knowledge, this is the first demonstration of the feasibility of deploying camera recording systems and developing machine learning-based workflow analysis solutions for open surgery, particularly in orthopaedics.