| Retinal image is an important tool for diagnosing the retinal diseases.Retinal structures and lesions can be observed directly in the retinal image.The changes of retinal structures and lesions indicate fundus diseases,such as glaucoma,diabetic retinopathy and so on.Therefore,retinal image processing and analysis is significant to early detection,diagnosis,treatment,auxiliary diagnosis and treatment of various retinal diseases.This thesis focuses on retinal structrues detection,retinal disease diagnosis and retinal image enhancement.The main works are summarized as follows:Automatic optic disc(OD)detection is an essential step for automatic screening of eye diseases.An OD localization method is proposed in this paper,which aims to locate OD robustly in retinal image with pathological changes.There are mainly three steps in this approach: region-of-interest(ROI)detection,candidate pixel detection,and confidence score calculation.The features of vessel direction,intensity,OD edges,and size of bright regions were extracted and employed in the proposed OD locating approach.Compared with the OD locating method based on vessel direction only,the proposed method could handle the following cases better: OD partially appears in retinal image,retinal vessels are not obvious in retinal image,or there are bright lesions in retinal images.Four public databases with total 340 retinal images were tested to evaluate the performance of our method.The proposed method can achieve an accuracy of 100%,95.8%,99.2%,97.8% for DRIVE database,STARE database,DIARETDB0 database,and DIARETDB1 database respectively.Comparison studies showed that the proposed approach is especially robust in the retinal images with diseases.A method to evaluate blurriness for cataract diagnosis in retinal images with vitreous opacity is proposed in this paper.Three types of features are extracted,which includes pixel number of visible structures,mean contrast between vessels and background,and local standard deviation.To avoid wrong detection of vitreous opacity as retinal structures,a morphological method is proposed to detect and remove such lesions from retinal visible structures segmentation.Based on the extracted features,a decision tree is trained to classify retinal images into five grades of cataract.The proposed approach was tested using 1355 clinical retinal images and the accuracies of two-class classification and five-grade grading compared with manual grading are 92.8% and 81.1%,respectively.The Kappa value between automatic grading and manual grading is 0.74 in five-grade grading,in which both variance and P-value are less than 0.001.Experimental results show that the grading difference between automatic grading with manual grading is all within 1 grade,which is much improvement compared with other available methods.The proposed grading method provides a universal measure of cataract severity and can facilitate the decision of cataract surgery.The good quality of color retinal image is essential for doctors to make a reliable diagnose in clinics.Due to the acquisition process or retinal diseases,poor illuminance,blur and low contrast are common in retinal images,which will seriously affect diagnosis.Image formation model of scattering is proposed to enhance color retinal images in this paper.Two parameters of this model,background illuminance and transmission map,are estimated based on extracted background and foreground.The complex nature of the foreground of a retinal image,involving pixels with both low and high intensity,posed a challenge to the proper extraction of these pixels.Therefore,a new method combining Mahalanobis distance discrimination and global spatial entropy-based contrast enhancement is proposed to extract foreground pixels.It extracts background and foreground in high intensity region and low intensity region respectively and it can perform well in blurry image with tiny intensity range.The proposed method is evaluated using 319 color retinal images from three different databases.Experimental results indicated that the proposed method can perform well on illumination problems,contrast enhancement and color preservation.This study proposes a new method of enhancing overall retinal image and produces better enhancement images than several state-of-the-art algorithms,especially for blurry retinal images. |
