A Distributed Image Feature Extraction Algorithm Suitable for Collaborative Learning

Abstract

The invention discloses a distributed image feature extraction algorithm suitable for collaborative learning. The algorithm is based on Radon transform and singular value decomposition. According to the geometric characteristics of Radon transform, a new invariant texture feature extraction based on Radon transform and wavelet transform is constructed. method, first extract the edge contour of the original image according to the principle of wavelet modulus maxima, construct the central moment of Radon transform for the edge image, obtain translation invariance, and then construct the scale invariant moment according to the statistical characteristics of Radon transform on the basis of the central moment , and finally, find the descending singular value vector of the even moment of the scale invariant moment, the singular value eigenvector has translation, scale and rotation invariance. The invention enhances the anti-noise ability while maintaining translation, rotation and scale invariance; and overcomes the influence of noise on image shape.

Description

technical field [0001] The invention relates to an image feature extraction algorithm, in particular to a distributed image feature extraction algorithm suitable for collaborative learning. Background technique [0002] Modern distance education is based on the Internet and terminals to realize teaching activities. The real-time remote virtual environment in teaching mainly includes distance classroom teaching, tutoring and answering questions, group study and discussion, etc. It is necessary to realize barrier-free and smooth teaching between teachers and students in the form of interaction and learning, and collaborative learning is an important way to improve the effectiveness of distance learning. However, the main factor restricting the development and application of collaborative learning is that the supporting collaborative learning system (CSCL) has not yet studied in depth the key technologies such as collaborative learning model, collaborative learning perception, ...

AI Technical Summary

Problems solved by technology

However, the main factor restricting the development and application of collaborative learning is that the supporting collaborative learning system (CSCL) has not yet studied in depth the key technologies such as collaborative learning model, collaborative learning perception, personalized copy consistency, and concurrency control based on collaborative learning semantics. In order to achieve sufficiently efficient, natural interactive and personalized learning

It also has the following disadvantages: 1. It is suitable for most of the current image retrieval algorithms in the field of collaborative image design, and is not suitable for the field of distance education collaborative learning

2. No retrieval scheme for complex objects in virtual real-time collaborative learning

However, this method uses HU invariant moments to describe the shape features of the image, which is sensitive to noise, because when the wavelet modulus maximum is used to extract the edge, there is still a certain degree of noise information near the edge due to the different selection of the threshold.

Teh's comprehensive analysis of several moments shows that Zernike moments have less noise sensitivity, but still cannot meet the needs of actual shape feature extraction

Because the Radon transform has a strong anti-noise ability, it has been used in image recognition. However, there are currently literatures that construct a Radon invariant that uses cyclic translation to obtain rotation invariance, which has a certain complexity, and the scale invariance construction more cumbersome

Images