Method for detecting character elements in natural scene image

Abstract

The invention relates to a method for detecting character elements in a natural scene image, and belongs to the field of computers. According to the method, firstly, a maximum pooling operation is used for performing non-maximum suppression on a target, then feature extraction is performed through a feature extraction network, and finally features of a top layer and a bottom layer are integrated through feature fusion, so that the representation capability of the features is improved, and the accurate position of the output character target is predicted in combination with a prediction network. According to the method of the invention, through designing a novel neural network structure, features of a character top layer and a character bottom layer in a natural scene image can be integrated, and accurate detection of character positions in horizontal arrangement, inclined arrangement and bending arrangement under a complex background can be achieved.

Description

technical field [0001] The invention belongs to the field of computers and relates to a method for detecting character elements in natural scene images. Background technique [0002] With the advancement of computer vision technology, text recognition technology has rapidly emerged and flourished. Its application has spread to every corner of the national economy and social life, and it is bringing about tremendous changes in human production methods, working methods and even life styles. With the deepening of applications, the market has higher and higher requirements for text detection methods in natural scenes with more complex scenes. [0003] Text detection is the prelude to subsequent operations such as text recognition, and its detection effect directly affects the effect of subsequent operations such as recognition. Text in natural scenes usually contains a lot of background interference, image lighting effects and other issues, and the aspect ratio of different lan...

AI Technical Summary

Problems solved by technology

Traditional detection methods are often unable to deal with many problems such as background interference. Deep learning methods can achieve better detection results, but deep learning methods rely on large data set training. The existing single-character data sets for text detection in natural scenes are due to the cost of labeling. Larger, so the data set is less, which will have a certain negative impact on algorithm development and effect testing

Images