Attention image segmentation method and device and medium
A technology of image segmentation and attention, applied in the field of image processing, to achieve the effect of improving efficiency, improving accuracy, and improving segmentation accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0078] The embodiment of the present invention provides a kind of attention image segmentation method, please refer to figure 1 , including the following steps:
[0079] S100, training the attention network, the attention network is a trainable coarse segmentation network; in this embodiment, the attention network is based on the ResNet network, but is not limited to this type of network; includes the original image to be segmented, and the backbone network structure ;
[0080] S110, the backbone network convolves the image through the convolution kernel, and extracts the feature map of the image; sets the step size of the convolution, and controls the size of the feature map after convolution through the step size of the convolution; in the backbone network, each time After one convolution, the size of the feature map of the image will be doubled. For example, the previous image was 200*200, and after one convolution, it becomes a 100*100 image;
[0081] S120, performing mu...
Embodiment 2
[0126] The embodiment of the present invention also provides an attention image segmentation system, please refer to Figure 6 , comprising: an extraction module, a fusion module, a first segmentation module, a transformation module and a second segmentation module;
[0127] The extraction module is used to convolve the image through the convolution kernel and extract several feature maps of the image;
[0128] The fusion module is used to select several feature maps and fuse them to obtain a fusion feature map;
[0129] The first segmentation module is used to obtain the first segmentation result of the image through the attention network and the fusion feature map;
[0130] The transformation module is used to select a segmentation network, and is used to perform size transformation on the first segmentation result of the image to obtain region information;
[0131] The second segmentation module is configured to perform weighted fusion of the image through the segmentatio...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More