A Time-Series Sketch Recognition Method Combining Texture Features and Shape Features

Abstract

The invention belongs to the field of computer vision, and discloses a time-series sketch recognition method that combines texture features and shape features. First, an image sequence is acquired according to the stroke order of the sketch; second, the texture features and shape features of each image are extracted to form a feature sequence corresponding to the image sequence; then, the features are input into a network consisting of two stages for learning, The two cyclic neural networks in the first stage accept the texture feature and shape feature of the image respectively. In the second stage, the output of the previous stage is first fused, and then input to the third cyclic neural network, and finally the result is obtained through the classifier. In this way, according to The order in the sequence is learned iteratively. Advantages: The geometric descriptor is used for sketch recognition, and the cyclic neural network is used to effectively learn the timing features of sketches, thereby significantly improving the defect that the original recognition model ignores the shape features and timing features of sketches, and improving the sketch recognition rate. .

Description

technical field [0001] The invention belongs to the field of computer vision, relates to image classification tasks, in particular to a time-series sketch recognition method which combines texture and shape features. Background technique [0002] Sketches that reflect the main features of an object are an effective way for people to communicate ideas. However, it is a huge challenge to automatically recognize hand-drawn sketches with variable stroke order and different styles, so that the automatic recognition rate has reached the ceiling in recent years. Most of the current methods, especially those based on deep networks, do not ignore the geometric features of sketches, and use the texture features that have achieved great success in natural image recognition for sketch recognition, while ignoring that texture features are not very important in sketches. Abundant defects; it is to treat sketches as handwritten letters with a fixed structural order, ignoring the timing of...

AI Technical Summary

Problems solved by technology

However, it is a huge challenge to automatically recognize hand-drawn sketches with variable stroke order and different styles, so that the automatic recognition rate has reached the ceiling in recent years

Most of the current methods, especially those based on deep networks, do not ignore the geometric features of sketches, and use the texture features that have achieved great success in natural image recognition for sketch recognition, while ignoring that texture features are not very important in sketches. Abundant defects; it is to treat sketches as handwritten letters with a fixed structural order, ignoring the timing of sketches that are different from other shapes, or failing to solve the disturbance of similar sketches with different hand-drawn orders to the results.

Images