Mode training method based on ensemble learning and mode indentifying method

Abstract

The invention provides a mode training method based on ensemble learning and a mode indentifying method. The mode training method comprises the following steps of: 1) carrying out dictionary learning on training samples to generate a redundant dictionary; 2) utilizing the redundant dictionary to carry out sparse encoding on the training samples to obtain a sparse encoding coefficient of each training sample; 3) carrying out sparse subspace division on all the training samples according to the sparse encoding coefficients; and 4) carrying out sub-model training on the training sample in each sparse subspace to obtain a sub-model for classifying. According to the mode training method based on the ensemble learning and the mode indentifying method, provided by the invention, higher indentifying performance can be obtained; and meanwhile, the training efficiency and the detection efficiency can be obviously improved.

Description

technical field [0001] The present invention relates to the field of intelligent systems, and more particularly, to the fields of pattern recognition and machine learning. Background technique [0002] For data with high dimensionality, the detection and recognition models trained by traditional methods on small data sets are difficult to cover all possible sample situations, and the generalization performance and detection accuracy are poor on open data sets. Especially for the rapidly growing image and video data on the Internet, not only has high feature dimension, but also has the characteristics of wide coverage, diverse content, and fast update. This kind of possible sample can improve the detection accuracy of the algorithm on the open multimedia data set. [0003] In 2009, M.Enzweiler and D.M.Gavrila published the article "Monocular Pedestrian Detection: Survey and Experiments" on pages 2179-2195 of IEEE Transactions on Pattern Analysis and Machine Intelligence thro...

AI Technical Summary

Problems solved by technology

[0003] In 2009, M.Enzweiler and D.M.Gavrila published the article "Monocular Pedestrian Detection: Survey and Experiments" on pages 2179-2195 of IEEE Transactions on Pattern Analysis and Machine Intelligence through the study of pedestrian detection in images. The combination of feature and pattern classifiers, the benefits brought by the increase in the training sample set are not obvious

However, this kind of weight coefficient cannot guarantee its sparsity, so the efficiency of training and fusion needs to be further improved, and the number of non-zero coefficients of sparse representation is small

Images