AUTHOR=Bai Yun , Li Jing , Shi Lianjun , Jiang Qin , Yan Biao , Wang Zhenhua 

TITLE=DME-DeepLabV3+: a lightweight model for diabetic macular edema extraction based on DeepLabV3+ architecture

JOURNAL=Frontiers in Medicine

VOLUME=Volume 10 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2023.1150295

DOI=10.3389/fmed.2023.1150295

ISSN=2296-858X

ABSTRACT=Diabetic macular edema (DME) is a complication of diabetes and a major cause of vision impairment in the patients with diabetes. To achieve high-efficiency and high-precision extraction of DME in OCT images, we proposed a lightweight model for DME extraction using a DeepLabV3+ (DME-DeepLabV3+) architecture. In the proposed model, MobileNetV2 model was used as the backbone for extracting the low-level features of DME. The improved ASPP with sawtooth wave-like dilation rate was used for extracting the high-level features of DME. Then, the decoder was used to fuse and refine the low-level and high-level features of DME. Finally, 1711 OCT images were collected from the Kermany dataset and the Affiliated Eye Hospital, Nanjing Medical University. 1369, 171, and 171 OCT images were randomly selected for training, validation, and testing, respectively. DME extraction performance of DME-DeepLabV3+ model was evaluated by comparing against other end-to-end models, including DeepLabV3+, FCN, U-Net, PSPNet, ICNet, and DANet. The result showed that DME-DeepLabV3+ model had better DME extraction performance than other models as shown by greater pixel accuracy (PA), mean pixel accuracy (MPA), precision (Pre), recall (Re), F1-score (F1), and mean Intersection over Union (MIoU), which were 98.71%, 95.23%, 91.19%, 91.12%, 91.15%, and 91.18%, respectively. Collectively, DME-DeepLabV3+ model is suitable for the extraction of DME and can assist the ophthalmologists in the management of ocular diseases.