青光眼视网膜神经纤维层的数码摄影及模糊影像的图像处理

目的 视网膜神经纤维层（ｒｅｔｉｎａｌ ｎｅｒｖｅ ｆｉｂｅｒ ｌａｙｅｒ,ＲＮＦＬ）的检测是青光眼早期诊断的重要手段,但照片冲洗难度大,部分患者难以显示清晰的ＲＮＦＬ图像,为些进行ＲＮＦＬ的数码摄影及模糊影像图像处理的研究。方法 采用Ｎｉｋｏｎ眼底摄像机,连接高分辨Ｋｏｄａｋ数码摄像机。图像处理采用Ｐｈｏｔｏｓｈｏｐ软件。对首都医科大学附属北京同仁医院眼科１０５例青光眼患者,３２例白内障患者,２

Digital photography and blurred image processing of retinal nerve fiber layer in glaucoma

OBJECTIVE: Red-free photographs are effective in detecting retinal nerve fiber layer (RNFL). The reasons why it is not used by most clinics are difficulty in laborious developing and blurred image evaluating. To account for getting real-time image and blurred image evaluating, the digital photography and computerized image processing are used. METHODS: RNFL photographs were taken by Nikon Fundus Camera connected with a high resolution digital camera (Kodak) and a green filter. As usual it is necessary to see the texture of RNFL in evaluating its image, but blurred images due to media opacities greatly lose their resolution. The way we used no focus on seeing the texture of RNFL, but on evaluating the distribution pattern of RNFL. RNFL images have more white reflex in temporal superior, temporal inferior, nasal superior and nasal inferior sections around optic disc according to the thickness of RNFL. Threshold segmentation was used, we shift a threshold bar to the position until the nasal or temporal section around optic disc is black with the software Photoshop. RESULTS: In the normal RNFL image, the white reflex was of butterfly pattern, i.e., the temporal superior and temporal inferior sections around optic disc were white. In cases with glaucoma, the characteristic of early damage of optic nerve was the loss of RNFL at the temporal superior and temporal inferior part of the retina, thus the normal butterfly pattern was changed, the temporal superior or/and temporal inferior sections were black. With this method it is more effective to evaluate RNFL. However, in cases with retinal pigment epithelium faint in color and in cases with choroidal defect, it is still difficult to evaluate the distribution of RNFL. CONCLUSIONS: It is an easy and effective way to detect glaucomatous damage with the digital photography and RNFL pattern evaluation.