Abstract
Diabetic Retinopathy (DR) is a complication of diabetes that can lead to blindness if not readily discovered. Automated screening algorithms have the potential to improve identification of patients who need further medical attention. However, the identification of lesions must be accurate to be useful for clinical application. The bag-of-visual-words (BoVW) algorithm employs a maximum-margin classifier in a flexible framework that is able to detect the most common DR-related lesions such as microaneurysms, cotton-wool spots and hard exudates. BoVW allows to bypass the need for pre- and post-processing of the retinographic images, as well as the need of specific ad hoc techniques for identification of each type of lesion. An extensive evaluation of the BoVW model, using three large retinograph datasets (DR1, DR2 and Messidor) with different resolution and collected by different healthcare personnel, was performed. The results demonstrate that the BoVW classification approach can identify different lesions within an image without having to utilize different algorithms for each lesion reducing processing time and providing a more flexible diagnostic system. Our BoVW scheme is based on sparse low-level feature detection with a Speeded-Up Robust Features (SURF) local descriptor, and mid-level features based on semi-soft coding with max pooling. The best BoVW representation for retinal image classification was an area under the receiver operating characteristic curve (AUC-ROC) of 97.8% (exudates) and 93.5% (red lesions), applying a cross-dataset validation protocol. To assess the accuracy for detecting cases that require referral within one year, the sparse extraction technique associated with semi-soft coding and max pooling obtained an AUC of 94.22.0%, outperforming current methods. Those results indicate that, for retinal image classification tasks in clinical practice, BoVW is equal and, in some instances, surpasses results obtained using dense detection (widely believed to be the best choice in many vision problems) for the low-level descriptors.
Highlights
For progressive diseases, such as complications of diabetes mellitus, early diagnosis has a huge impact on prognosis, allowing corrective or palliative measures before irreversible organ damage takes place
We believe that the good performance of dense features on Messidor is due to the presence of very challenging images
Automated lesion detection has a huge potential to facilitate the identification of diabetic retinopathy progression, and the access to care, for rural and remote communities, providing a screening tool able to determine which patients need to be referred to specialists
Summary
For progressive diseases, such as complications of diabetes mellitus, early diagnosis has a huge impact on prognosis, allowing corrective or palliative measures before irreversible organ damage takes place. In the case of Diabetic Retinopathy (DR), early detection is crucial to prevent vision loss. Screening patients for early signs of DR pathology is important to prevent the disease or limit its progression. In disfavored, rural or isolated communities, the access to healthcare professionals – to ophthalmology specialists – is difficult or not possible, reducing opportunities for early detection and timely treatment of DR. In order to be useful, the automated system must identify a specific type of lesions that occurs both in isolation and in combination with other types of lesions, and make accurate decisions on the need to refer the patient to a specialist for further assessment
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
