Neural network-based automatic image annotation method, system, device and medium
An automatic image and neural network technology, applied in the fields of computer vision and artificial intelligence, which can solve the problems of lack of prediction of the number of labels, failure to consider the relationship between labels and labels, and low label accuracy.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0102] Example one, such as figure 1 As shown, an automatic image annotation method based on neural network includes the following steps:
[0103] S1: Obtain an experimental data set, and use a pre-trained convolutional neural network model to extract image features of the experimental data set;
[0104] S2: Obtain the image to be labeled from the test set of the experimental data set, and according to the image characteristics, in the training set of the experimental data set, use the k-nearest neighbor method to calculate the neighborhood image set and A first label field corresponding to the neighborhood image set;
[0105] S3: Constructing a tag semantic association model between the first tag domain and the second tag domain corresponding to the training set, and calculating in the second tag domain according to the tag semantic association model and the first tag domain A third label domain associated with each first label in a label domain;
[0106] S4: Calculate the similarit...
Example Embodiment
[0158] Embodiment two, such as Figure 5 As shown, an automatic image annotation system based on neural network includes an acquisition module, an extraction module, a calculation module, and an annotation module:
[0159] The acquisition module is used to acquire an experimental data set;
[0160] The extraction module is used to extract image features of the experimental data set by using a pre-trained convolutional neural network model;
[0161] The acquisition module is also used to acquire the image to be labeled from the test set of the experimental data set;
[0162] The calculation module is used to calculate the neighborhood image set of the image to be labeled and the first neighborhood image set corresponding to the neighborhood image set in the training set of the experimental data set according to the image feature. Label domain
[0163] The calculation module is also used to construct a tag semantic association model between the first tag domain and the second tag domain ...
Example Embodiment
[0172] Embodiment 3. Based on Embodiment 1 and Embodiment 2, this embodiment also discloses a neural network-based automatic image labeling device, including a processor, a memory, and stored in the memory and capable of running on the processor The computer program on the figure 1 The specific steps from S1 to S5 are shown.
[0173] Through the computer program stored in the memory and running on the processor, the automatic image annotation of the present invention is realized. Based on the convolutional neural network, the relationship between the image and the image, the relationship between the image and the label, and the relationship between the label and the label are fully considered. Relations, combining similarity and probability models to predict the target label of the image to be annotated, the prediction accuracy has been significantly improved, thus greatly improving the accuracy of the annotation, making the effect of automatic image annotation better, and better ...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap