Method, system and device for predicting infant brain age based on T2-weighted imaging

Abstract

The present invention relates to the field of image analysis technology, and specifically to a method, system, and device for predicting infant brain age based on T2-weighted imaging. To solve the technical problem of low infant brain age prediction accuracy in the prior art, the present invention first extracts age-related local brain features and brain myelin features at different scales based on T2-weighted imaging to obtain multi-scale local features and multi-scale myelin features; then, the multi-scale local features and multi-scale myelin features are element-wise added at corresponding scales to enhance their complementarity and representation capabilities in the brain age prediction task, and after obtaining multi-scale myelin enhancement features, adaptive multi-branch feature fusion processing is performed to obtain multi-scale brain fusion features; the multi-scale brain fusion features are element-wise multiplied with channel attention information of deep myelin features to obtain whole-brain enhancement features; finally, the whole-brain enhancement features are linearly and nonlinearly processed to obtain infant brain age prediction results.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and in particular to a method, system and device for predicting infant brain age based on T2-weighted imaging. Background Art

[0002] Structural magnetic resonance imaging (MRI), as a non-invasive imaging technology, can provide high-resolution image information of brain anatomical structures. The richness and high quality of MRI data support the development of brain age prediction algorithms, giving them significant advantages in early infant brain development research, abnormal development screening and intervention strategies.

[0003] Currently, convolutional neural networks (CNNs) are often used to extract MRI features and predict brain age. This also includes the use of image information from single-modality or multi-modality neuroimaging technologies to predict brain age. However, these technologies have weak ability to express the image features obtained from infant brain images, and the accuracy of infant brain age prediction results is low. Some studies have focused on the fact that myelin information in neuroimaging is an important biological feature of rapid development in infancy. Myelin information can be used as a neurodevelopmental indicator to predict infant brain age. However, myelin information currently often relies on doctors to manually outline the myelin area. This process is not only time-consuming and labor-intensive, but also results in poor final brain age prediction and unstable prediction quality. Summary of the Invention

[0004] The purpose of the present invention is to provide a method, system and device for predicting infant brain age based on T2-weighted imaging.

[0005] The technical solutions of the present invention are as follows:

[0006] A method for predicting infant brain age based on T2-weighted imaging includes the following operations:

[0007] S1. Based on T2-weighted imaging, obtain whole-brain images and white matter images; obtain local features of different scales of the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically: downsample and dense feature stacking are performed on the whole-brain image, repeat the downsampling and dense feature stacking process twice, and repeat the downsampling and dense feature stacking process three times to obtain shallow local features, middle local features, and deep local features; obtain myelin sheath information of different scales of the white matter image to obtain shallow myelin sheath features, middle myelin sheath features, and deep myelin sheath features;

[0008] S2. Add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features element-by-element to obtain shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features, respectively; the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are fused through adaptive multi-branch features to obtain multi-scale brain fusion features; the deep myelin features are processed through convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features are multiplied element-by-element with the multi-scale brain fusion features to obtain whole-brain enhancement features;

[0009] S3. The whole-brain enhanced features are processed linearly and nonlinearly to obtain the infant brain age prediction results.

[0010] The specific operations for obtaining shallow myelin features in S1 are as follows: the white matter image is processed with local details enhancement in different directions to obtain multi-directional myelin detail enhancement features; the multi-directional myelin detail enhancement features are subjected to convolution and average pooling to obtain shallow myelin features.

[0011] The specific operations for obtaining the middle layer myelin sheath features in S1 are as follows: the shallow layer myelin sheath features are processed with spatial attention to obtain the myelin sheath spatial features; the myelin sheath spatial features and the shallow layer myelin sheath features are multiplied element by element, and then convolution and average pooling are performed to obtain the middle layer myelin sheath features.

[0012] The deep myelin features in S1 are obtained by performing local detail enhancement, convolution and average pooling on the middle myelin features in different directions.

[0013] The specific operations of the adaptive multi-branch feature fusion processing in S2 are as follows: the shallow myelin enhancement features, the middle myelin enhancement features and the deep myelin enhancement features are respectively subjected to maximum pooling and depthwise separable convolution, and then spliced to obtain multi-scale brain fusion features.

[0014] The specific operations for obtaining the whole-brain image and white matter image in S1 are as follows: T2-weighted imaging is subjected to voxel normalization and skull information stripping to obtain the initial whole-brain image; the initial whole-brain image is subjected to white matter segmentation to obtain the initial white matter image; the initial whole-brain image is subjected to bias correction and then aligned with the initial white matter image to obtain the whole-brain image and white matter image.

[0015] The specific operations of channel attention processing in S1 are as follows: the deep myelin sheath features are convolved to obtain deep myelin sheath convolution features; the deep myelin sheath convolution features are averaged and pooled to obtain deep myelin sheath pooled features; the deep myelin sheath pooled features are fully connected and then multiplied element-wise with the deep myelin sheath pooled features to obtain the initial myelin sheath attention features; the initial myelin sheath attention features are processed nonlinearly to obtain the myelin channel attention features.

[0016] A system for predicting infant brain age based on T2-weighted imaging, used to implement the above-mentioned method for predicting infant brain age based on T2-weighted imaging, comprising:

[0017] The multi-scale local feature and multi-scale myelin feature generation modules are used to obtain whole-brain images and white matter images based on T2-weighted imaging; obtain local features of different scales in the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically, the whole-brain image is downsampled and densely stacked, the downsampling and dense feature stacking processes are repeated twice, and the downsampling and dense feature stacking processes are repeated three times to obtain shallow local features, middle local features, and deep local features; myelin information of different scales in the white matter image is obtained to obtain shallow myelin features, middle myelin features, and deep myelin features;

[0018] The whole-brain enhancement feature generation module is used to add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features element by element, respectively, to obtain the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features; the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are processed by adaptive multi-branch feature fusion to obtain multi-scale brain fusion features; the deep myelin features are processed by convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features are multiplied element by element with the multi-scale brain fusion features to obtain the whole-brain enhancement features;

[0019] The infant brain age prediction result generation module is used to obtain the infant brain age prediction result by linear and nonlinear processing of the whole brain enhancement features.

[0020] A device for predicting infant brain age based on T2-weighted imaging comprises a processor and a memory, wherein the processor implements the above-mentioned method for predicting infant brain age based on T2-weighted imaging when executing a computer program stored in the memory.

[0021] A computer-readable storage medium is used to store a computer program, wherein when the computer program is executed by a processor, it implements the above-mentioned method for predicting infant brain age based on T2-weighted imaging.

[0022] The beneficial effects of the present invention are:

[0023] The present invention provides a method for predicting infant brain age based on T2-weighted imaging. First, based on T2-weighted imaging, whole-brain images and white matter images that can respectively reflect the global structure and myelination characteristics of the whole brain are obtained, and age-related local brain features and brain myelin features at different scales are extracted to obtain multi-scale local features and multi-scale myelin features; then, the multi-scale local features and multi-scale myelin features are added element-by-element at corresponding scales, and the myelination information and age-related features in the whole-brain channel are integrated to enhance the complementarity and representation ability of the two in the brain age prediction task. After obtaining the multi-scale myelin enhancement features, adaptive multi-brain feature fusion processing is performed to obtain the multi-scale brain fusion features; the multi-scale brain fusion features are multiplied element-by-element with the channel attention information of the deep myelin features to obtain the whole-brain enhancement features; finally, the whole-brain enhancement features are linearly and nonlinearly processed to obtain the infant brain age prediction results. DETAILED DESCRIPTION

[0024] This embodiment provides a method for predicting infant brain age based on T2-weighted imaging, including the following operations:

[0025] S1. Based on T2-weighted imaging, obtain whole-brain images and white matter images; obtain local features of different scales of the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically: downsample and dense feature stacking are performed on the whole-brain image, repeat the downsampling and dense feature stacking process twice, and repeat the downsampling and dense feature stacking process three times to obtain shallow local features, middle local features, and deep local features; obtain myelin sheath information of different scales of the white matter image to obtain shallow myelin sheath features, middle myelin sheath features, and deep myelin sheath features;

[0026] S2. Add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features element-by-element to obtain shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features, respectively; the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are fused through adaptive multi-branch features to obtain multi-scale brain fusion features; the deep myelin features are processed through convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features are multiplied element-by-element with the multi-scale brain fusion features to obtain whole-brain enhancement features;

[0027] S3. The whole-brain enhanced features are processed linearly and nonlinearly to obtain the infant brain age prediction results.

[0028] S1. Based on T2-weighted imaging, whole-brain images and white matter images are obtained; local features of whole-brain images at different scales are obtained, and shallow local features, middle local features, and deep local features are obtained; myelin information of white matter images at different scales is obtained, and shallow myelin features, middle myelin features, and deep myelin features are obtained.

[0029] Based on T2-weighted imaging, whole-brain images and white matter images are obtained, which can reflect the global structure and myelination characteristics of the whole brain respectively. Age-related local brain features and brain myelin features at different scales are extracted to obtain multi-scale local features and multi-scale myelin features.

[0030] First, whole-brain images and white matter images were obtained based on T2-weighted imaging. Specifically, the original data collected at the hospital were T2-weighted images in the Digital Imaging and Communications in Medicine (DICOM) format. The T2-weighted images were converted to three-dimensional images in the Neuroimaging Informatics Technology Initiative (NIFTI) format, and patient privacy data was deleted. The T2-weighted images were then voxel-normalized to 1 × 1 × 1 mm. The skull was then stripped using FSL software to obtain an initial whole-brain image. The initial whole-brain image was then subjected to white matter segmentation using FSL software to ensure accurate extraction of white matter areas, resulting in an initial white matter image. The initial whole-brain image was then bias-corrected to eliminate low-frequency brightness unevenness in the image and registered with the initial white matter image. Both images were registered to the UNC Infant 0-1-2 atlas and cropped to 150 × 198 × 156 to ensure spatial consistency, resulting in the whole-brain and white matter images.

[0031] Then, local features at different scales of the whole-brain image are obtained, and age-related structural features at different scales are extracted. This enriches the whole-brain local features to obtain shallow, mid-level, and deep-level local features. Specifically, the whole-brain image is downsampled and densely stacked, repeated twice and three times, to achieve feature reuse and information transfer at different levels, enhance the expressive power of local features, and obtain shallow, mid-level, and deep-level local features. Dense feature stacking includes, but is not limited to, implementation via a dense convolutional network (DenseNet), and downsampling includes, but is not limited to, implementation via convolutional blocks.

[0032] At the same time, myelin information at different scales of white matter images is obtained to enrich myelin features and obtain shallow myelin features, middle myelin features and deep myelin features.

[0033] The shallow myelin features are obtained by performing local detail enhancement on the white matter image in different directions to obtain multi-directional myelin detail enhancement features. These multi-directional myelin detail enhancement features are then convolved and average pooled to obtain shallow myelin features. Local detail enhancement in different directions is achieved by performing conventional convolution, center difference convolution, horizontal difference convolution, and vertical difference convolution on the white matter image, followed by element-by-element addition. This emphasizes the central pixel features, horizontal edge features, and vertical edge features of the white matter image, thereby enhancing the details of the myelin region in the white matter image.

[0034] The specific operations for obtaining the middle-layer myelin features are as follows: the shallow myelin features are processed with spatial attention to learn the characteristic distribution of the myelin region, thereby obtaining the myelin spatial features. The myelin spatial features are then multiplied element-wise with the shallow myelin features, followed by convolution and average pooling. This automatically focuses on the details of the myelin sheaths in the white matter, strengthens the representation of the myelin region, and obtains the middle-layer myelin features. The specific operations for spatial attention processing are as follows: the shallow myelin features are processed with average pooling and maximum pooling, followed by channel-wise concatenation, three-dimensional convolution, and sigmoid function processing to obtain the myelin spatial features.

[0035] The deep myelin features are obtained by performing local detail enhancement, convolution and average pooling on the middle myelin features in different directions.

[0036] S2. Add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features element by element, respectively, to obtain shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features; the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are processed by adaptive multi-branch feature fusion to obtain multi-scale brain fusion features; the deep myelin features are processed by convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features and the multi-scale brain fusion features are multiplied element by element to obtain the whole-brain enhancement features.

[0037] The multi-scale local features and multi-scale myelin features are added element-by-element at the corresponding scales, and the myelination information is integrated with the age-related features in the whole-brain channel to enhance their complementarity and representation ability in the brain age prediction task. After obtaining the multi-scale myelin enhancement features, adaptive multi-branch feature fusion processing is performed to obtain the multi-scale brain fusion features; the multi-scale brain fusion features are multiplied element-by-element with the channel attention information of the deep myelin features to obtain the whole-brain enhancement features.

[0038] First, the shallow local features, middle local features, and deep local features are added element-by-element with the shallow myelin features, middle myelin features, and deep myelin features, respectively, to achieve the fusion of local features and myelin features of the corresponding scale brain regions, and integrate the myelination information extracted from the white matter myelin channel with the age-related features in the whole-brain channel to enhance their complementarity and representation ability in the brain age prediction task, and obtain shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features that reflect the local detail features and myelin information of brain regions at different levels, respectively.

[0039] Then, the shallow, middle, and deep myelin enhancement features are processed through adaptive multi-branch feature fusion. Based on the extraction of features at different scales, multi-scale information fusion is achieved, resulting in a multi-scale brain fusion feature with rich features and prominent details. The specific operation of the adaptive multi-branch feature fusion process is as follows: the shallow, middle, and deep myelin enhancement features are respectively subjected to maximum pooling and depthwise separable convolution, and then spliced to obtain the multi-scale brain fusion feature. In depthwise separable convolution processing, the input is convolved channel by channel, focusing only on the spatial information within each channel, and then the output of each channel is integrated through point-by-point convolution.

[0040] At the same time, the deep myelin features undergo convolution and channel attention processing to better focus on myelin information, resulting in myelin channel attention features. The channel attention processing operation is as follows: the deep myelin features undergo convolution processing to obtain deep myelin convolution features; the deep myelin convolution features undergo average pooling to obtain deep myelin pooling features; the deep myelin pooling features undergo fully connected processing (consisting of two fully connected layers) and are then element-wise multiplied with the deep myelin pooling features to obtain the initial myelin attention features; the initial myelin attention features undergo nonlinear processing (achievable through a sigmoid function) to obtain the myelin channel attention features.

[0041] Finally, the myelin channel attention features and the multi-scale brain fusion features are multiplied element-by-element to further enhance the complementarity and representation ability of the whole-brain structural information and myelin information, and obtain the whole-brain enhanced features.

[0042] S3. The whole-brain enhanced features are processed linearly and nonlinearly to obtain the infant brain age prediction results.

[0043] The whole-brain enhanced features are processed linearly and nonlinearly (which can be achieved through the Linear function and the Softmax function respectively) to obtain the infant brain age prediction results.

[0044] To verify the effectiveness of the method of this embodiment, the following experiments were conducted.

[0045] Experimental Setup: The dataset collected at the hospital was randomly divided into training, test, and validation sets in an 8:1:1 ratio. The experimental development environment used was PyTorch 2.1.0. Training was performed on an NVIDIA RTX 3090 GPU. The model was trained for 100 epochs using the Adam optimizer with an initial learning rate of 0.000 and a batch size of 5. The model was trained for approximately 12 hours. The experimental environment and specific settings are shown in Table 1.

[0046] Table 1 Summary of experimental parameter settings

[0047]

[0048] The experimental results are shown in Table 2. It can be seen that the prediction method of this embodiment has a small error in predicting the infant's brain age and a high prediction accuracy.

[0049] Table 2 Prediction results of infant brain age in this example

[0050]

[0051] This embodiment further provides a system for predicting infant brain age based on T2-weighted imaging, which is used to implement the above-mentioned method for predicting infant brain age based on T2-weighted imaging, including:

[0052] The multi-scale local feature and multi-scale myelin feature generation modules are used to obtain whole-brain images and white matter images based on T2-weighted imaging; obtain local features of different scales in the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically, the whole-brain image is downsampled and densely stacked, the downsampling and dense feature stacking processes are repeated twice, and the downsampling and dense feature stacking processes are repeated three times to obtain shallow local features, middle local features, and deep local features; myelin information of different scales in the white matter image is obtained to obtain shallow myelin features, middle myelin features, and deep myelin features;

[0053] The whole-brain enhancement feature generation module is used to add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features element by element, respectively, to obtain the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features; the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are processed by adaptive multi-branch feature fusion to obtain multi-scale brain fusion features; the deep myelin features are processed by convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features are multiplied element by element with the multi-scale brain fusion features to obtain the whole-brain enhancement features;

[0054] The infant brain age prediction result generation module is used to obtain the infant brain age prediction result by linear and nonlinear processing of the whole brain enhancement features.

[0055] This embodiment also provides a device for predicting infant brain age based on T2-weighted imaging, comprising a processor and a memory, wherein the processor implements the above-mentioned method for predicting infant brain age based on T2-weighted imaging when executing a computer program stored in the memory.

[0056] This embodiment further provides a computer-readable storage medium for storing a computer program, wherein when the computer program is executed by a processor, the computer program implements the above-mentioned method for predicting infant brain age based on T2-weighted imaging.

[0057] This embodiment provides a method for predicting infant brain age based on T2-weighted imaging. First, based on T2-weighted imaging, whole-brain images and white matter images that can respectively reflect the global structure and myelination characteristics of the whole brain are obtained, and age-related local brain features and brain myelin features at different scales are extracted to obtain multi-scale local features and multi-scale myelin features. Then, the multi-scale local features and multi-scale myelin features are element-wise added at corresponding scales, and the myelination information is integrated with the age-related features in the whole-brain channel to enhance the complementarity and representation ability of the two in the brain age prediction task. After obtaining the multi-scale myelin enhancement features, adaptive multi-brain feature fusion processing is performed to obtain the multi-scale brain fusion features. The multi-scale brain fusion features are element-wise multiplied with the channel attention information of the deep myelin features to obtain the whole-brain enhancement features. Finally, the whole-brain enhancement features are linearly and nonlinearly processed to obtain the infant brain age prediction results.

[0058] This embodiment provides a method for predicting infant brain age based on T2-weighted imaging. By collaboratively extracting myelin information from whole-brain images and white matter images, the method significantly improves the accuracy and robustness of prediction. The fusion mechanism effectively integrates myelination characteristics with age-related information, providing a reliable biomarker for early neurodevelopmental assessment in infants and has important clinical application potential.

Claims

1. A method for predicting infant brain age based on T2-weighted imaging, characterized in that: The following operations are included: S1, based on T2-weighted imaging, obtain whole brain images and white matter images; Obtaining local features of different scales of the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically, downsampling and dense feature stacking processing are performed on the whole-brain image, the downsampling and dense feature stacking processing is repeated twice, and the downsampling and dense feature stacking processing is repeated three times to obtain shallow local features, middle local features, and deep local features; Obtain myelin sheath information at different scales of white matter images to obtain superficial myelin sheath features, middle myelin sheath features, and deep myelin sheath features; S2. Add the shallow local features, middle local features, and deep local features to the shallow myelin sheath features, middle myelin sheath features, and deep myelin sheath features, element by element, to obtain the shallow myelin sheath enhancement features, middle myelin sheath enhancement features, and deep myelin sheath enhancement features; The superficial myelin enhancement features, middle myelin enhancement features and deep myelin enhancement features are processed by adaptive multi-branch feature fusion to obtain multi-scale brain fusion features; The deep myelin features are processed by convolution and channel attention to obtain the myelin channel attention features; the myelin channel attention features are multiplied element-wise with the multi-scale brain fusion features to obtain the whole brain enhancement features; S3. The whole-brain enhanced features are processed linearly and nonlinearly to obtain the infant brain age prediction results.

2. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: The operation of obtaining the superficial myelin sheath characteristics in S1 is specifically as follows: The white matter image is processed with local detail enhancement in different directions to obtain multi-directional myelin detail enhancement features; the multi-directional myelin detail enhancement features are subjected to convolution and average pooling to obtain shallow myelin features.

3. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: The specific operation of obtaining the middle layer myelin sheath characteristics in S1 is: The shallow myelin features are processed with spatial attention to obtain the myelin spatial features; the myelin spatial features and the shallow myelin features are multiplied element by element, and then convolution and average pooling are performed to obtain the middle myelin features.

4. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: In S1, the deep myelin sheath features are obtained by performing local detail enhancement processing, convolution and average pooling processing on the middle myelin sheath features in different directions.

5. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: In S2, the operations of the adaptive multi-branch feature fusion processing are specifically as follows: The shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are respectively subjected to maximum pooling and depth-wise separable convolution, and then spliced to obtain multi-scale brain fusion features.

6. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: In S1, the operations for obtaining the whole brain image and the white matter image are specifically as follows: The T2-weighted image was voxel-normalized and skull information was stripped to obtain the initial whole-brain image. The initial whole-brain image was processed for white matter segmentation to obtain the initial white matter image. The initial whole-brain image was processed for bias correction and then registered with the initial white matter image to obtain the whole-brain image and white matter image.

7. The method for predicting infant brain age based on T2-weighted imaging according to claim 1, wherein: In S1, the channel attention processing operation is specifically as follows: The deep myelin sheath features are processed by convolution to obtain deep myelin sheath convolution features; the deep myelin sheath convolution features are processed by average pooling to obtain deep myelin sheath pooling features; the deep myelin sheath pooling features are processed by full connection and then multiplied element-by-element with the deep myelin sheath pooling features to obtain the initial myelin sheath attention features; the initial myelin sheath attention features are processed by nonlinearity to obtain the myelin channel attention features.

8. A system for predicting infant brain age based on T2-weighted imaging, used to implement the method for predicting infant brain age based on T2-weighted imaging according to claim 1, characterized in that: include: The multi-scale local feature and multi-scale myelin feature generation modules are used to obtain whole-brain images and white matter images based on T2-weighted imaging; obtain local features of different scales in the whole-brain image to obtain shallow local features, middle local features, and deep local features; specifically, the whole-brain image is downsampled and densely stacked, the downsampling and dense feature stacking processes are repeated twice, and the downsampling and dense feature stacking processes are repeated three times to obtain shallow local features, middle local features, and deep local features; myelin information of different scales in the white matter image is obtained to obtain shallow myelin features, middle myelin features, and deep myelin features; The whole-brain enhancement feature generation module is used to add the shallow local features, middle local features, and deep local features to the shallow myelin features, middle myelin features, and deep myelin features, respectively, to obtain the shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features; The shallow myelin enhancement features, middle myelin enhancement features, and deep myelin enhancement features are processed by adaptive multi-branch feature fusion to obtain multi-scale brain fusion features; the deep myelin features are processed by convolution and channel attention to obtain myelin channel attention features; the myelin channel attention features and the multi-scale brain fusion features are multiplied element-wise to obtain the whole brain enhancement features; The infant brain age prediction result generation module is used to obtain the infant brain age prediction result by linear and nonlinear processing of the whole brain enhancement features.

9. A device for predicting infant brain age based on T2-weighted imaging, characterized in that: The method comprises a processor and a memory, wherein when the processor executes the computer program stored in the memory, the method for predicting infant brain age based on T2-weighted imaging as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that Used to store a computer program, wherein when the computer program is executed by a processor, the method for predicting infant brain age based on T2-weighted imaging according to any one of claims 1 to 7 is implemented.

Images