Vector diagram retrieval method and system based on deep learning

Abstract

The invention discloses a vector diagram retrieval method and system based on deep learning. The retrieval method comprises the steps: generating target bitmaps in one-to-one correspondence with target vector diagrams based on the target vector diagrams; combining the target vector diagram and the target bitmap to form a plurality of file streams with identifiers; obtaining feature information of the target bitmap and the to-be-retrieved image based on deep learning; comparing the feature information of the target bitmap and the feature information of the to-be-retrieved image, and preliminarily screening out a plurality of target images related to the to-be-retrieved image from the target bitmap; and according to the feature difference between the vector diagram and the bitmap, carrying out relevancy sorting on the target image. According to the method, repeated drawing of the user vector diagram is reduced, a training model is not needed, rapid deployment can be achieved, and the effect that a client can accurately retrieve the server-side vector diagram and download the server-side vector diagram to the client side by using the bitmap in a short time is achieved.

Description

technical field [0001] The invention belongs to the field of image retrieval, in particular to a vector image retrieval method and system based on deep learning. Background technique [0002] The current mainstream image retrieval method is to use convolutional neural network to extract image features, use metric learning methods such as Euclidean distance to calculate the distance of image features, sort image distances, obtain primary retrieval results, and then use contextual information and flow information of image data to obtain primary retrieval results. The shape structure is used to reorder the image retrieval results, so as to improve the accuracy of image retrieval and obtain the final retrieval results. [0003] The drawbacks of such methods when retrieving vector graphics are: [0004] 1. Time-consuming: the reordering process adds additional processing time; [0005] 2. Poor retrieval accuracy: For the retrieval of images of the same category (mechanical draw...

AI Technical Summary

Problems solved by technology

[0004] 1. Time-consuming: the reordering process adds additional processing time;

[0005] 2. Poor retrieval accuracy: For the retrieval of images of the same category (mechanical drawing, etc.) with small intra-category differences, and when there are large differences between vector graphics and bitmaps in terms of line thickness, position offset, noise, scaling, etc., directly use This method is difficult to obtain the target image of interest;

[0006] 3. Difficult to label: Difficult to classify vector graphics

Images