Ophthalmology OCT attenuation coefficient calculation method for backscattering compensation

Abstract

The invention relates to the technical field of optical coherence tomography (OCT), and particularly discloses a method for compensating backscattering of a soft glass film of a retina layer. An SS-OCT imaging system is used for collecting an image, an original OCT image is obtained, a retinal pigment epithelium (RPE) layer and a Drusen area below the RPE layer are obtained through image enhancement and noise filtering, an initial optical attenuation coefficient (OAC) and an initial backscattering coefficient alpha of the RPE layer and the Drusen area are calculated through a local depth resolution (LDR) algorithm, backscattering compensation is carried out on the Drusen area, and therefore the RPE layer and the Drusen area are obtained. The backscattering coefficient of the Drusen region is matched with the RPE layer, and the OAC of the compensated RPE layer and the Drusen region is calculated by using a local depth resolution (LDR) algorithm. According to the method, the calculation precision of the optical attenuation coefficient can be remarkably improved, and finally more reliable data support is provided for AMD diagnosis.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical coherence tomography (OCT), and is particularly directed to a method for accurately calculating the light attenuation coefficient in ophthalmic applications. The method mainly solves the problem that in the calculation of the light attenuation coefficient based on OCT, the backscattering and absorption characteristics of different fundus tissues are different, which affects the accuracy of the calculation and the diagnostic effect. Background Art

[0002] Optical coherence tomography (OCT) is a non-invasive imaging technology based on the principle of low-coherence light interference. It can provide high-resolution cross-sectional images of the retina and is widely used in the diagnosis and research of retinal diseases. Since OCT can show the changes in the layered structure of the retina in detail, it has become an important tool in clinical ophthalmology and retinal disease detection.

[0003] Age-related macular degeneration (AMD) is one of the main causes of severe vision loss, especially in the elderly. AMD is mainly divided into two types: wet and dry. Dry AMD can further develop into geographic atrophy (GA), which is caused by the degeneration of the retinal pigment epithelium (RPE). Damage and loss of function of the RPE are considered to be one of the main pathological characteristics of GA.

[0004] During the pathological development of AMD, the decline in the function of the RPE layer has a direct impact on the health of the retina. The RPE has multiple roles in maintaining retinal function. When the RPE function is impaired, its support for outer segment photoreceptors is reduced, resulting in a gradual loss of photoreceptor function and ultimately triggering the formation of GA.

[0005] Changes in the RPE are a key factor in the development of GA, and OCT provides a non-invasive means of evaluating the decline of the RPE layer. For early detection and progression monitoring of GA, OCT can evaluate the state of the RPE by calculating the optical attenuation coefficient (OAC) within the retinal layer. The OAC value is the degree to which light is attenuated due to absorption and scattering during propagation in the tissue, reflecting the optical properties of the tissue. Especially for patients with dry AMD, the OAC value of the RPE layer is an important parameter for detecting the degree of atrophy.

[0006] However, in AMD patients, the calculation of OAC is challenged by the presence of multiple subretinal Drusen. Drusen is a subretinal deposit formed by the accumulation of lipids and proteins, which has the characteristics of strong absorption and weak scattering. This special optical property affects the accurate calculation of OAC, making it impossible to truly reflect the attenuation characteristics of the RPE layer.

[0007] One of the current assumptions for accurately calculating OAC is that it relies on the complete attenuation of light in the tissue, that is, the light signal is completely attenuated in the depth direction. However, this assumption is not completely true in AMD patients, especially in the presence of multiple Drusen. Existing methods can effectively deal with the problem of complete attenuation of light in retinal tissue, but for AMD patients, especially in the case of strong absorption and weak scattering caused by Drusen, it cannot accurately reflect the actual optical attenuation characteristics.

[0008] Another common assumption is that the ratio of light absorption to scattering (backscatter coefficient α) is the same in different tissues. However, in the AMD-lesioned fundus, this assumption is no longer valid due to the significant differences in the optical properties of each tissue. In particular, the presence of Drusen makes the assumption of α inaccurate, leading to misjudgment of OAC. This inaccuracy seriously affects the assessment of the degree of RPE decline, especially the detection of early stages of GA. Therefore, in order to improve the accuracy of AMD diagnosis, a new method is needed to overcome the limitations of existing methods in the assumption of the ratio of scattering and absorption of different tissues.

[0009] This patent proposes a backscatter compensation ophthalmic OCT attenuation coefficient calculation method, focusing on calculating the backscatter attenuation ratio (α) of the retinal layer to accurately calculate the value of OAC. Compared with the traditional method, this method overcomes the assumption that the absorption and scattering ratios are the same, especially for the Drusen lesion area where AMD is common, by correcting the calculation of the α value, the accuracy of OAC calculation is improved. Summary of the invention

[0010] This method mainly solves the problem that in the calculation of the light attenuation coefficient based on OCT, due to the different backscattering and absorption characteristics of different fundus tissues, the calculation accuracy and diagnostic effect are affected. To this end, the present invention proposes a Drusen backscatter compensation method, the core steps of which are as follows:

[0011] A. Preprocessing of retinal images: Preprocess the retinal OCT images using image enhancement and noise filtering techniques to preliminarily segment the boundaries of RPE and Drusen to improve the accuracy of subsequent calculations.

[0012] B. Initial calculation of backscatter coefficient: Calculate the initial backscatter coefficient of the RPE layer and Drusen, and analyze the scattering characteristics of the two to determine the scattering difference that needs to be compensated.

[0013] The backscattering coefficient α of the RPE layer is calculated by formula (1): R ;

[0014]

[0015] The backscatter coefficient α of the Drusen region is calculated by formula (2): D .

[0016]

[0017] C. Backscatter compensation of soft glass membrane: The backscatter of the soft glass membrane is adjusted using a compensation algorithm to match the backscatter characteristics of the RPE layer. The algorithm can adaptively adjust the compensation parameters to cope with individual differences among different patients.

[0018] The signal after Drusen compensation is calculated by formula (3):

[0019]

[0020] D. OAC calculation of unified tissue: The compensated soft glass membrane and RPE layer are regarded as a unified tissue, and the optical attenuation coefficient of the tissue is recalculated using the improved LDR algorithm, and the corrected result is output.

[0021] The unified tissue OAC is calculated by formula (4).

[0022] BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 Diagram of the steps for calculating the light attenuation coefficient. Figure 2 (a) is the OCT structure of the RPE layer and Drusen; (b) is the schematic diagram of the RPE layer and Drusen; (c) is the distribution diagram of Drusen; (d) is the division of the normal RPE tissue area and the Drusen area. The area outside the red curve is the non-pigmented area. The area inside the black curve represents the pigmented area. Figure 3The results of LDR algorithm combined with backscatter compensation are shown: (a) OCT enface image of the RPE layer of AMD patients. (b) is the B-scan image at the orange dashed line position in (a). (c) OAC result obtained using the traditional DR algorithm. (d) OAC result obtained using the fitting-based DR algorithm. (e) OCT B-scan image after drusen backscatter compensation. (f) OAC result of RPE obtained using the method proposed in this paper. The red arrow points to the abnormal signal area, the yellow arrow indicates the area with weak OCT signal intensity, the blue arrow points to the drusen area, and the brown arrow indicates the erroneous signal caused by fitting. DETAILED DESCRIPTION

[0024] See also Figure 1 The specific implementation process is as follows:

[0025] In step S1, the OCT image is preprocessed to improve the boundary segmentation accuracy of the RPE layer and the soft glass membrane through image enhancement and noise filtering.

[0026] In step S2, the initial backscattering coefficients of the RPE and the soft glass membrane Drusen are calculated.

[0027] In step S3, a compensation algorithm is applied to adjust the backscattering characteristics of the soft glass film to match the characteristics of the RPE layer.

[0028] In step S4, the compensated soft glass membrane and the RPE layer are taken as a whole, and the optical attenuation coefficient of the whole tissue is calculated using an improved LDR algorithm.

Claims

1. An OCT attenuation imaging method for backscatter compensation in ophthalmology, characterized in that: The details are as follows: Obtaining optical coherence tomography (OCT) images of retinal tissue; Image enhancement and noise filtering techniques were used to preliminarily extract the boundary between RPE and soft glaucoma (Drusen); The optical attenuation coefficient (OAC) and backscattering coefficient of the RPE layer and Drusen area were calculated; Perform backscatter compensation on the Drusen area to make the backscatter coefficient of the Drusen area match the RPE layer; The compensated OAC was calculated using the local depth resolution (LDR) algorithm.

2. The method according to claim 1, characterized in that The steps to identify the RPE layer and Drusen area include: The upper boundary of the RPE layer, the upper boundary of Drusen and the Bruch's membrane (BM) were marked based on the structural features of the OCT images.

3. The method according to claim 2, characterized in that The backscatter coefficient of the RPE layer and Drusen area is calculated, including the following steps: The backscattering coefficient α of the RPE layer is calculated by formula (1): R ; The backscatter coefficient α of the Drusen region is calculated by formula (2): D .

4. The method according to claim 3, characterized in that Backscatter compensation in the Drusen region includes: According to formula (3), using α R and α D The signal in the Drusen area is compensated. It can be obtained by following the steps below:

5. The method according to claim 4, characterized in that After compensating the Drusen area, the RPE layer and compensated Drusen were considered as a whole tissue with the same backscatter coefficient to improve the calculation accuracy of OAC. The calculation formula (4) of OAC can be obtained by the following steps:

Images