Target classifying method based on local and depth feature assembling

Abstract

The invention provides a target classifying method based on a local and depth feature assembling. The method mainly includes the steps of depth convolution feature extracting, local feature extracting, encoding and assembling through classifiers, wherein in the process, Fisher vectors of descriptors and scale invariant feature transformation (SIFT) descriptors are extracted in a final complete connection layer of a depth network, a support vector machine (SVM) is trained for each feature with the Fisher vectors as the encoding strategies, the assembling through the classifiers is trained and tested, an input data set is optimized and classified, and finally voting is conducted and a final decision is made. By means of an intermediate layer of the depth network, the classification capacityof the features obtained from the complete connection layer can be enhanced, one independent classifier is trained for each feature, and therefore good classification performance is achieved; meanwhile, computation cost is low, and various applications of the target classification technology can be easily obtained.

Description

technical field [0001] The invention relates to the field of object classification, in particular to an object classification method based on local and deep feature sets. Background technique [0002] Object recognition and classification are current research hotspots in the fields of computer vision and artificial intelligence. It is an important branch of pattern recognition technology, which can be defined as the process of processing, analyzing, describing, classifying and interpreting vector information representing objects. Recognized objects include text, sound, images, etc. Target recognition and classification can be applied to fingerprint recognition and face recognition for identity confirmation, license plate recognition in intelligent traffic management, seed recognition in agriculture, food quality detection technology and electrocardiogram recognition technology in medicine, etc., and can be further developed Extended to text and voice recognition, remote sen...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problems of poor classification effect and high computational cost, the purpose of the present invention is to provide a target classification method based on local and deep feature sets, first extracting descriptors and scale-invariant feature transformations from the last fully connected layer of the deep network (SIFT) the Fisher vector of the descriptor, then train a support vector machine (SVM) for each feature using the Fisher vector as an encoding strategy, and then train and test the classifier ensemble to optimize the input dataset Classify, and finally vote and arrive at the final decision

Images