The aim of this study is to assess the performance two machine-learning technologies, namely, deep learning (DL) and support vector machine (SVM) algorithms, for detecting central retinal vein occlusion (CRVO) in ultrawide-field fundus images. Images from 125 CRVO patients (<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M1"><mml:mrow><mml:mi>n</mml:mi><mml:mo>=</mml:mo><mml:mn>125</mml:mn></mml:mrow></mml:math> images) 202 non-CRVO normal subjects...

To evaluate the accuracy of detecting glaucoma visual field defect severity using deep-learning (DL) classifier with an ultrawide-field scanning laser ophthalmoscope.One eye 982 open-angle (OAG) patients and 417 healthy eyes were enrolled. We categorized into 3 groups according to damage (Humphrey Field Analyzer 24-2 program) [early; -6 dB (mean deviation) or better, moderate; between -12 dB, severe as mean deviation worse]. In total, 558 images (446 for training 112 grading) from early OAG...

To investigate and compare the efficacy of two machine-learning technologies with deep-learning (DL) support vector machine (SVM) for detection branch retinal vein occlusion (BRVO) using ultrawide-field fundus images.This study included 237 images from 236 patients BRVO a mean±standard deviation age 66.3±10.6y 229 176 non-BRVO healthy subjects mean 64.9±9.4y. Training was conducted deep convolutional neural network to construct DL model. The sensitivity, specificity, positive predictive...

Purpose: To evaluate the ability of deep learning (DL) models to detect obstructive meibomian gland dysfunction (MGD) using in vivo laser confocal microscopy images. Methods: For this study, we included 137 images from individuals with MGD (mean age, 49.9 ± 17.7 years; 44 men and 93 women) 84 normal glands 53.3 19.6 29 55 women). We constructed trained 9 different network structures used single ensemble DL calculated area under curve, sensitivity, specificity compare diagnostic abilities DL....

Evaluating the discrimination ability of a deep convolution neural network for ultrawide-field pseudocolor imaging and autofluorescence retinitis pigmentosa. In total, 373 images (150, pigmentosa; 223, normal) obtained from patients who visited Department Ophthalmology, Tsukazaki Hospital were used. Training with convolutional on these learning data objects was conducted. We examined K -fold cross validation ( = 5). The mean area under curve group 0.998 (95% confidence interval (CI)...

Abstract The present study aimed to conduct a real-time automatic analysis of two important surgical phases, which are continuous curvilinear capsulorrhexis (CCC), nuclear extraction, and three other phases cataract surgery using artificial intelligence technology. A total 303 cases registered in the clinical database Ophthalmology Department Tsukazaki Hospital were used as dataset. Surgical videos downsampled resolution 299 × 168 at 1 FPS image each frame. Next, based on start end times...

Purpose: To evaluate the efficacy of deep learning in judging need for rebubbling after Descemet's endothelial membrane keratoplasty (DMEK). Methods: This retrospective study included eyes that underwent DMEK (rebubbling group: RB group) and same number did not require (non-RB group), based on medical records. classify group, randomly selected images from anterior segment optical coherence tomography at postoperative day 5 were evaluated by corneal specialists. The criterion was condition...

Conjunctival hyperaemia is a common clinical ophthalmological finding and can be symptom of various ocular disorders. Although several severity classification criteria have been proposed, none include objective criteria. Neural networks deep learning utilised in ophthalmology, but not for the purpose classifying conjunctival objectively. To develop grading software, we used 3700 images as training data 923 validation test data. We trained nine neural network models validated performance...

The present study aims to describe the use of machine learning (ML) in predicting occurrence postoperative refraction after cataract surgery and compares accuracy this method conventional intraocular lens (IOL) power calculation formulas. In total, 3331 eyes from 2010 patients were assessed. objects divided into training data test data. constants for IOL formulas model ML optimized using Then, was predicted formulas, or models calculated We evaluated SRK/T formula, Haigis Holladay 1 Hoffer Q...

The present study aimed to compare the accuracy of diabetic retinopathy (DR) staging with a deep convolutional neural network (DCNN) using two different types fundus cameras and composite images. included 491 ultra-wide-field ophthalmoscopy optical coherence tomography angiography (OCTA) images that passed an image-quality review were graded as no apparent DR (NDR; 169 images), mild nonproliferative (NPDR; 76 moderate NPDR (54 severe (90 proliferative (PDR; 102 images) by three retinal...

Abstract This study was performed to estimate choroidal thickness by fundus photography, based on image processing and deep learning. Colour photography central examinations were in 200 normal eyes with serous chorioretinopathy (CSC). Choroidal under the fovea measured using optical coherence tomography images. The adaptive binarisation method used delineate vessels within colour photographs. Correlation coefficients calculated between vascular density (defined as vasculature appearance...

Surgical skill levels of young ophthalmologists tend to be instinctively judged by in practice, and hence a stable evaluation is not always made for single ophthalmologist. Although it has been said that standardizing presents difficulty as surgical methods vary greatly, approaches based on machine learning seem promising this objective. In study, we propose method displaying the information necessary quantify techniques cataract surgery real-time. The proposed consists two steps. First, use...

Exon-skipping is a powerful genetic tool, especially when delivering genes using an AAV-mediated full-length gene supplementation strategy difficult owing to large length of genes. Here, we used engineered human induced pluripotent stem cells and artificial intelligence evaluate clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR associated protein 9-based exon-skipping vectors targeting the retinal pigment epithelium (RPE). The model system was choroideremia; this...

The article reports on two children with relapsing pancreatitis. In both cases operative cholangiogram demonstrated unusually long common pancreatico-biliary channel not associated cystic dilatation of the bile duct. possibility and hypothetical mechanism that caused pancreatitis, are described. Exploratory laparotomy should be considered for having recurrent abdominal pain hyperamylasemia. For surgical treatment this type pancreatitis excision duct gallbladder hepaticoenterostomy will best method.

In this paper, a method of achieving high accuracy in discriminating right and left eyes is proposed for ophthalmic surgery. A VGG16 convolutional neural network employed to construct main classifier. The data presented the classifier are some frames sampled at regular intervals from surgery videos. Before classifying be examined, determines whether they suitable or not improve discrimination as possible. other words, causing erroneous omitted. determination depends on image characteristics...

Abstract In this paper, we propose a method of checking the eye lotion instillation for ophthalmology patients. Our first estimates tilt angles an dropper bottle from acceleration values measured by triaxial sensor attached to bottle. It next prepares data each which is equal sequence standardized values, as be presented discrimination model. employs either long short‐term memory (LSTM short) or bidirectional (B_LSTM construct Once present checked our model, it produces certainty degree...