Transfer Learning with Convolutional Neural Networks for Diabetic Retinopathy Image Classification. A Review

Abstract

Diabetic retinopathy (DR) is a dangerous eye condition that affects diabetic patients. Without early detection, it can affect the retina and may eventually cause permanent blindness. The early diagnosis of DR is crucial for its treatment. However, the diagnosis of DR is a very difficult process that requires an experienced ophthalmologist. A breakthrough in the field of artificial intelligence called deep learning can help in giving the ophthalmologist a second opinion regarding the classification of the DR by using an autonomous classifier. To accurately train a deep learning model to classify DR, an enormous number of images is required, and this is an important limitation in the DR domain. Transfer learning is a technique that can help in overcoming the scarcity of images. The main idea that is exploited by transfer learning is that a deep learning architecture, previously trained on non-medical images, can be fine-tuned to suit the DR dataset. This paper reviews research papers that focus on DR classification by using transfer learning to present the best existing methods to address this problem. This review can help future researchers to find out existing transfer learning methods to address the DR classification task and to show their differences in terms of performance.

A Comprehensive Study of Machine Learning Methods on Diabetic Retinopathy Classification

“…Instead, weights are imported from another CNN, which is trained on a large dataset. The ImageNet dataset is the most famous dataset for transferring weights of trained models [6].…”

“…The suitability of transfer learning for DR classification can be evaluated by comparing a model that was trained from scratch, to its fine-tuned version. Several studies (such as Masood et al [39]; Wan et al [5]; Xu et al [40]) agree that using transfer learning increases the accuracy of a model significantly in the classification of DR [6]. A vast number of images are needed to sufficiently train a deep learning model.…”

“…InceptionV3 is one of the most used models in DR classification studies. The high performance of InceptionV3 may be associated with the inception module, which was found to be very useful in DR images [6].…”

Designing and Manufacturing of Automatic Robotic Lawn Mower

“…Object cutting technology is a condition-oriented and basic image processing technology, it can quickly cut the large range of images that may contain objects under detection to increase efficiency [19]. Image recognition technology has been continuously improved and has been successfully applied in many industries, such as digital rights management (DRM) [20], robotic arm path planning [21], studies on the determination of fruit ripeness [1,4], lighting control technology [22], and diabetic retinopathy diagnosis [23][24][25][26]. With the investment of related resources and the evolution of technology, more innovative applications will be developed in the future.…”

Hierarchical Transfer Learning for Multilingual, Multi-Speaker, and Style Transfer DNN-Based TTS on Low-Resource Languages

“…Especially for deep learning, a study by [26] classifies deep transfer learning (DTL) into different four categories: instances-based DTL, mapping-based DTL, network-based DTL, and adversarial-based DTL. Some DNN-based systems have successfully applied deep transfer learning, including TTS [31], image classification [34] [35], machine translation [36] [37], automatic speech recognition [38] [39] [40], language identification [41], and sentiment classification [42].…”