Complex wavelet based quality assessment for AS-OCT images with application to Angle Closure Glaucoma diagnosis. Issue 130 (July 2016)

• Record Type: Journal Article
• Authors: Niwas, Swamidoss Issac; Jakhetiya, Vinit; Lin, Weisi; Kwoh, Chee Keong; Sng, Chelvin C.; Aquino, Maria Cecilia; Victor, Koh; Chew, Paul T.K.
• Is Part Of: Computer methods and programs in biomedicine. Issue 130(2016)
• Page Start: 13
• Page End: 21
• Publication Date: 2016-07
• Subjects: Optical coherence tomography -- Angle closure glaucoma -- Complex wavelets -- Local binary pattern -- Image quality assessment -- Machine learning; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2016.03.011
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: A new quality assessment method for AS-OCT images using wavelet based LBP features. It is a first work; so far there is no objective assessment of AS-OCT image quality. The proposed algorithm does not require any additional information from the AS-OCT device. This work aimed at collecting high quality AS-OCT images for Glaucoma diagnosis. Our proposed quality index score is similar to the quality assessment of Glaucoma experts. Abstract: Background and objectives: Angle closure disease in the eye can be detected using time-domain Anterior Segment Optical Coherence Tomography (AS-OCT). The Anterior Chamber (AC) characteristics can be quantified from AS-OCT image, which is dependent on the image quality at the image acquisition stage. To date, to the best of our knowledge there are no objective or automated subjective measurements to assess the quality of AS-OCT images. Methods: To address AS-OCT image quality assessment issue, we define a method for objective assessment of AS-OCT images using complex wavelet based local binary pattern features. These features are pooled using the Naïve Bayes classifier to obtain the final quality parameter. To evaluate the proposed method, a subjective assessment has been performed by clinical AS-OCT experts, who graded the quality of AS-OCT images on a scale of good, fair, and poor. This was done based on the ability to identify the AC structures including the position of the scleral spur. Results: We compared the results of the proposed objective assessment with the subjective assessments. From this comparison, it is validated that the proposed objective assessment has the ability of differentiating the good and fair quality AS-OCT images for glaucoma diagnosis from the poor quality AS-OCT images. Conclusions: This proposed algorithm is an automated approach to evaluate the AS-OCT images with the intention for collecting of high quality data for further medical diagnosis. Our proposed quality index has the ability of automatic objective and quantitative assessment of AS-OCT image quality and this quality index is similar to glaucoma specialist.