Automated Analysis for Retinopathy of Prematurity by Deep Neural Networks

Abstract

Retinopathy of Prematurity (ROP) is a retinal vasproliferative disorder disease principally observed in infants born prematurely with low birth weight. ROP is an important cause of childhood blindness. Although automatic or semiautomatic diagnosis of ROP has been conducted, most previous studies have focused on "plus" disease, which is indicated by abnormalities of retinal vasculature. Few studies have reported methods for identifying the "stage" of ROP disease. Deep neural networks have achieved impressive results in many computer vision and medical image analysis problems, raising expectations that it might be a promising tool in automatic diagnosis of ROP. In this paper, convolutional neural networks (CNNs) with novel architecture is proposed to recognize the existence and severity of ROP disease per-examination. The severity of ROP is divided into mild and severe cases according to the disease progression. The proposed architecture consists of two sub-networks connected by a feature aggregate operator. The first sub-network is designed to extract high-level features from images of the fundus. These features from different images in an examination are fused by the aggregate operator, then used as the input for the second subnetwork to predict its class. A large dataset imaged by RetCam 3 is used to train and evaluate the model. The high classification accuracy in the experiment demonstrates the effectiveness of proposed architecture for recognizing ROP disease.

Artificial Intelligence in Retinopathy of Prematurity Diagnosis

“… 29 , 30 For example, a DL system called DeepROP achieved a sensitivity of 96.62% (95% confidence interval, 92.29%–98.89%) and a specificity of 99.32% (95% confidence interval, 96.29%–9.98%) for the detection of ROP (vs no ROP). 31 , 32 Zhao et al 30 reported the development of a DL system that can automatically draw the border of zone 1 on a fundus image as a diagnostic aid. Mulay et al 29 were the first to report the identification of a peripheral ROP ridge (stage) directly in a fundus image.…”

DeepEC: An error correction framework for dose prediction and organ segmentation using deep neural networks

“…Since 2012, when AlexNet 12 won the 2012 ILSVRC competition, 13 numerous important breakthroughs in computer vision have been achieved using DCNNs. [14][15][16] To automatically predict treatment planning for patients who need radiotherapy, and benefit from the development of DCNNs, previous studies used DCNNs to build a model to predict dose distribution. [2][3][4] The similarity of these studies is that they constructed dose prediction models based on the treatment planning of patients treated with IMRT.…”

Primena Algoritama Mašinskog Učenja U Oftamološkoj Analizi Slika Oka Kod Prevremeno Rođenih Beba

“…Na osnovu prethodnih radova na ovu temu [4,5,6], izabran je podskup metoda mašinskog učenja koje su dale najbolje rezultate. Odabrane su konvolutivne neuronske mreže za koje se smatra da su trenutno najperspektivnije za problem detekcije.…”