Fine-grained eye fundus image grading algorithm based on bilinear pooling and attention mechanism
A fundus image and attention technology, applied in the field of intelligent medical image processing, can solve problems such as sample imbalance, diabetic retinopathy image classification difficulties, etc., and achieve the effect of wide application scenarios
Pending Publication Date: 2021-06-22
JILIN UNIV
View PDF7 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] In order to solve the current difficulties in grading images of diabetic retinopathy and the unbalanced samples in the data set, the present invention provides a fine-grained fundus image grading algorithm based on bilinear pooling and attention mechanism
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0053] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings, but it is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be covered by the present invention. within the scope of protection.
[0054] The present invention provides a fine-grained fundus image classification algorithm based on bilinear pooling and attention mechanism, and the algorithm includes the following steps:
[0055] Step 1: Get the dataset
[0056] Download the dataset from Kaggle's Diabetic Retinopathy Detection Challenge (EyePACS).
[0057] Step 2: Dataset preprocessing
[0058] Use Opencv to adjust the size of the image obtained in step 1 to 512×512, denoise some overexposed images in the data set, and crop the useless parts of the image, so that the final image co...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a fine-grained fundus image grading algorithm based on bilinear pooling and an attention mechanism, and the algorithm comprises the steps: employing two groups of non-homologous CNNs to extract fundus data set image features, and employing the outputs of two groups of different CNN models as inputs; calculating a loss back propagation network by supplementing a cross entropy loss function, and finally testing the obtained network structure. The neural network model combining the bilinear pooling structure and the attention mechanism is a novel diabetic retinopathy fundus image grading and classifying method. The diabetic retinopathy fundus image grading detection algorithm based on the combination of the bilinear pooling model and the attention mechanism module is an accurate and efficient automatic detection and classification algorithm, has an extremely important value for clinic, and has a wide application scene.
Description
technical field [0001] The invention belongs to the field of intelligent medical image processing, and relates to a fine-grained fundus image classification algorithm based on bilinear pooling and attention mechanism. Background technique [0002] Deep learning is one of the latest trends in machine learning and artificial intelligence research. It is also one of the most popular scientific research trends today. Deep learning methods have revolutionized the fields of computer vision and machine learning. In recent years, deep learning methods have received extensive attention in medical image processing. For some specific tasks, deep learning methods have been shown to match or exceed the performance of medical experts. [0003] In fundus images, retinopathy lesions with microaneurysms, hemorrhages, hard exudates, cotton wool-like exudates, and neovascularization can be observed. According to the severity of fundus lesions, diabetic retinopathy can be divided into five ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06V40/10G06N3/045G06F18/24G06F18/214
Inventor 刘萍萍金百鑫杨晓康周求湛
Owner JILIN UNIV