PurposeDry eye disease affects hundreds of millions of people worldwide and is one of the most common causes for visits to eye care practitioners. The fluorescein tear breakup time test is currently widely used to diagnose dry eye disease, but it is an invasive and subjective method, thus resulting in variability in diagnostic results. This study aimed to develop an objective method to detect tear breakup using the convolutional neural networks on the tear film images taken by the non-invasive device KOWA DR-1α.MethodsThe image classification models for detecting characteristics of tear film images were constructed using transfer learning of the pre-trained ResNet50 model. The models were trained using a total of 9,089 image patches extracted from video data of 350 eyes of 178 subjects taken by the KOWA DR-1α. The trained models were evaluated based on the classification results for each class and overall accuracy of the test data in the six-fold cross validation. The performance of the tear breakup detection method using the models was evaluated by calculating the area under curve (AUC) of receiver operating characteristic, sensitivity, and specificity using the detection results of 13,471 frame images with breakup presence/absence labels.ResultsThe performance of the trained models was 92.3%, 83.4%, and 95.2% for accuracy, sensitivity, and specificity, respectively in classifying the test data into the tear breakup or non-breakup group. Our method using the trained models achieved an AUC of 0.898, a sensitivity of 84.3%, and a specificity of 83.3% in detecting tear breakup for a frame image.ConclusionsWe were able to develop a method to detect tear breakup on images taken by the KOWA DR-1α. This method could be applied to the clinical use of non-invasive and objective tear breakup time test.

According to the data from the World Health Organization, patients with visual function disorders, such as amblyopia and intermittent exotropia, have increased significantly and become an important factor contributing to the global burden of disease.Currently, existing research has found that many human visual function disorders can be reflected in abnormal brain activities without ocular damage. The main clinical methods currently rely on subjective psychophysical tests and ophthalmological examinations, and the detection and utilization of neural signals are insufficient. Moreover, when assessing patients with language expression disorders, including infants, or uncooperative subjects, these methods pose considerable challenges, thereby complicating the implementation of vision assessments and indicating certain limitations of traditional methods.Therefore, this paper introduces a steady-state visual-evoked EEG (S2VEEG) dataset, which includes 60 amblyopia patients, 14 intermittent exotropia patients, and 30 individuals with normal visual function. These patients were carefully diagnosed and selected by professional ophthalmologists in the hospital. A 64-channel EEG cap was used to collect the EEG data records of the subjects under steady-state visual stimulation. These physiological data can be used to explore new possible physiological indicators of visual function and for visual function assessment using emerging artificial intelligence technologies, thus promoting the development of the visual function disorder detection field.The paper extracts differential entropy and small-worldness property as features and uses SVM,RGCB and EmT classifiers for training. The results show that the recognition accuracy among amblyopia patients, intermittent exotropia patients, and individuals with normal visual function reaches 95.38%. This classification performance indicates significant differences among patients with different degrees of visual function disorders, demonstrating the good usability of the S2VEEG dataset and verifying the potential of small-worldness property as a biomarker for visual function disorder detection.

Performance of Working Optometrists versus Optometry Students.

Figure 1.A. Axial diffusion-weighted image revealed a restricted water diffusion lesion in the left medulla oblongata (red arrow). B. Magnetic resonance angiography disclosed no abnormality. (C) 3T-coronal HRMR views of T1-weighted image demonstrate vessel wall with strong smooth, concentric wall enhancement and thickening of the terminal segment of left BA. (D) 3T-Axial HRMR of T1-weighted image also showed eccentric wall thickening and mild enhancement of the M3 segment of the left MCA.

The paper deals with one approach to the objective assessment of an ultrasound examination, for which an application software suite using modern information technologies, mathematical methods, and artificial intelligence methods is proposed. A three-stage model of medical diagnostics is proposed and the concept of a predictor is introduced. The main aim of the study is to recognize closed asymptomatic contours in ultrasound images, to organize their monitoring. It is shown that a change in one of the properties of the predictor indicates the presence of possible dynamics in this area, which may be the onset of pathology.