Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Three-dimensional image segmentation method based on double-path attention coding and decoding network

A 3D image and attention technology, applied in image analysis, image enhancement, image data processing, etc., can solve problems such as lack of, limited semantic segmentation accuracy, and insufficient consideration of pixel-to-pixel connection, so as to reduce false positives and false positives. Counter-example, optimize image edge details, and improve the effect of expressive ability

Pending Publication Date: 2021-11-12
SHANGHAI UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Compared with FCN, U-Net has a more symmetrical encoding and decoding structure, and the skip connection from encoding to decoding part is helpful for the recovery of position information, but since the basic module of constructing the network structure is a simple convolution block, there are certain The degree of gradient disappearance limits the increase of network depth; in addition, U-Net does not fully consider the relationship between pixels and pixels, and lacks the exploration of dependencies between local features, thus affecting the accuracy of the final segmentation results.
Loss function is an important tool for optimizing network parameters. Losses such as cross entropy and similarity coefficient lack the ability to optimize the network to explore image boundary features, which limits the improvement of semantic segmentation accuracy.

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Three-dimensional image segmentation method based on double-path attention coding and decoding network
  • Three-dimensional image segmentation method based on double-path attention coding and decoding network
  • Three-dimensional image segmentation method based on double-path attention coding and decoding network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0068] see figure 1 , in this embodiment, a three-dimensional image segmentation method based on a two-way attention encoding and decoding network is provided. This method constructs an efficient two-way attention encoding and decoding network structure, uses boundary loss to optimize network parameters, and improves network Segmentation accuracy for 3D image data.

[0069] The method of the present invention uses certain three-dimensional medical images to train the model, obtains the model parameters of such data, and then obtains high-precision prediction of similar segmented data other than samples, and the method of the present invention includes the following steps:

[0070] (1) Randomly crop the original image used for training into smaller image blocks, preprocess the small image blocks to obtain a clearer image, and save the preprocessed data locally;

[0071] (2) Build a two-way attention-based encoding and decoding network, input the training set data into the netw...

Embodiment 2

[0076] This embodiment is basically the same as Embodiment 1, especially in that:

[0077] In this example, if figure 2 As shown, image preprocessing includes the following steps:

[0078] (1-1) cutting the three-dimensional image data into image blocks of 12×224×244 pixels;

[0079] (1-2) Determine whether the image block is a grayscale image, and use a normalization algorithm to convert the grayscale image for the non-grayscale image;

[0080] (1-3 Use Gaussian filtering to remove noise points in the image;

[0081] (1-4) Use histogram equalization to stretch the gray distribution of the image and enhance the contrast of the image;

[0082] (1-5) Use the Laplacian operator to realize image edge sharpening processing, enhance the gray level mutation in the image, that is, reduce the area where the gray level changes slowly;

[0083] (1-6) Divide and save the preprocessed image data.

[0084] In this embodiment, image preprocessing is performed, and the data is stored lo...

Embodiment 3

[0086] This embodiment is basically the same as Embodiment 2, and the special features are:

[0087] In this embodiment, the two-way attention codec network includes three sub-network modules, namely: (a) encoder network, (b) two-way attention network and (c) decoder network; using residual block, Maximum pooling, average pooling, and dual-path blocks build the encoder, and the encoder network construction includes the following steps:

[0088] (2-1-1) Use a residual block to construct the first layer of the encoder to adapt to the input of different data dimensions, and use the maximum pooling to reduce the dimensionality of the output of the first layer;

[0089] (2-1-2) Use two dual-path blocks in the second layer of the encoder to explore the low-level texture features of the image, and use maximum pooling to reduce the dimensionality of the output of the second layer;

[0090] (2-1-3) Use 3 dual-path blocks in the third layer of the encoder to explore the advanced abstra...

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

No PUM Login to View More

Abstract

The invention discloses a three-dimensional image segmentation method based on a double-path attention coding and decoding network. The method comprises the following steps: firstly, preprocessing a training image, then constructing a double-path attention coding and decoding network, optimizing parameters of the network by using boundary loss, and performing prediction segmentation on a three-dimensional image by using a trained model; and finally, adjusting the probability graph by using a dense conditional random field and a maximum connected region algorithm, optimizing a segmentation result, and storing an output post-processing result. According to the invention, a double-path attention module is integrated into the coding and decoding network, the segmentation precision of the network on the three-dimensional image is improved, the time cost and the labor cost of three-dimensional image processing are reduced, and the progress and development of the corresponding schools and production departments are promoted.

Description

technical field [0001] The invention relates to the field of computer vision three-dimensional image analysis and processing. Aiming at three-dimensional image data, a three-dimensional image segmentation method based on a two-way attention encoding and decoding network is proposed. The present invention can be applied to three-dimensional image segmentation in the fields of materials science and medicine, improves the accuracy of three-dimensional image segmentation, reduces the time cost and labor cost of three-dimensional image processing, and promotes the progress and development of corresponding academic and industrial circles. Background technique [0002] Image semantic segmentation is a common concern in image processing and other fields. Semantic segmentation is to allow the computer to segment according to the content of the image. Segmentation is to segment different objects in the image from the pixel level, label each pixel in the original image, and classify it...

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
no application Login to View More
IPC IPC(8): G06T7/11G06T7/187G06T5/00G06T5/40G06K9/62G06N3/04G06N3/08
CPCG06T7/11G06T7/187G06T5/40G06N3/08G06T2207/10012G06T2207/20081G06T2207/20084G06T2207/20132G06T2207/20221G06T2207/20024G06T2207/20048G06N3/045G06F18/253G06F18/214G06T5/73G06T5/90
Inventor 韩越兴李小龙钱权王冰
Owner SHANGHAI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com