Task model acquisition method and device, storage medium and electronic device
A task model and acquisition method technology, applied in the computer field, can solve problems such as poor model performance and slow convergence speed of model training, and achieve the effects of ensuring model performance, solving slow convergence speed of model training, and improving convergence speed
Active Publication Date: 2022-02-08
腾讯医疗健康(深圳)有限公司
View PDF6 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Embodiments of the present invention provide a task model acquisition method and device, a storage medium, and an electronic device, so as to at least solve the technical problems of slow model training convergence speed and poor model performance in the randomly initialized network model in the related art
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 approach
[0140] As an optional implementation, the input module includes:
[0141] (1) Extracting submodules, used to extract target image features from the first image using the first encoder;
[0142] (2) The processing sub-module is used to use the target pooling layer and the target fully connected layer to process the target image features to obtain target prediction information, wherein the target pooling layer and the target fully connected layer are located between the first encoder and the first Between decoders, the target prediction information is used to represent the predicted distribution of pixel values of each pixel in the second image over multiple pixel value intervals;
[0143] (3) An input sub-module, configured to input the target image features and target prediction information to the first decoder to obtain the first predicted image.
[0144] Through this embodiment, by setting an average pooling layer and a fully connected layer between the first encoder and ...
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 task model acquisition method and device, a storage medium and an electronic device. Wherein, the method includes: obtaining target model parameters obtained by performing model training on the first training model, the first training model includes: a first encoder and a first decoder, and the target model parameters include: the first encoder corresponding to the first A model parameter, the input of the first training model is the first image of the first channel of the target optical image, and the output of the first training model after training is the predicted image corresponding to the second image of the second channel of the target optical image , the graphic similarity between the predicted image and the second image is greater than or equal to the target similarity threshold; the target model parameters are used to initialize the model parameters of the second training model to obtain the initial task model, and the second training model includes the same structure as the first encoder The second encoder: using the target optical image to train the initial task model to obtain the target task model.
Description
technical field [0001] The present invention relates to the field of computers, in particular to a task model acquisition method and device, a storage medium and an electronic device. Background technique [0002] At present, when using a network model to perform optical image processing tasks, it is necessary to use sample optical images to perform model training on the initial network model to obtain a trained target network model. For example, network models such as CNN (Convolutional Neural Network, convolutional neural network), FCN (Fully Convolutional Network, fully convolutional network) can be used to perform tasks such as image classification and image segmentation. [0003] When performing model training, the network model is first assigned randomly generated initial model parameters (model parameters of the random initialization network model); the sample optical image is input into the randomly initialized network model, and the output results of the network mod...
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 Patent Type & Authority Patents(China)
IPC IPC(8): G06N3/08G06N3/04G16H30/00
CPCG06N3/082G16H30/00G06N3/045
Inventor 李悦翔郑冶枫
Owner 腾讯医疗健康(深圳)有限公司