Image recognition method and device

Abstract

Provided is an image identification method and device. The method comprises the following steps: obtaining an original image; scanning the original image through a window in a preset size to obtain a plurality of candidate regions; calculating image domain rectangular integral image features and feature values as well as image gradient domain rectangular integral image features and feature values of the current candidate region through a preset feature learning model; calculating feature value of the current candidate region through a plurality of strong classifiers arranged according to the sequence of complexity in a cascaded decision classifier, and according to the obtained image domain rectangular integral image features and feature values as well as the image gradient domain rectangular integral image features and feature values of the current candidate region; and when the obtained feature value of the current candidate region is larger than the threshold value of the corresponding strong classifier, determining the current candidate region is an image of a target object. According to the scheme above, image identification accuracy can be improved, operation amount is reduced, and image identification speed is improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to an image recognition method and device, and a mobile terminal. Background technique [0002] Image recognition refers to the technology of using computers to process, analyze and understand images to identify targets and objects in various patterns. Face recognition is a biometric technology for identification based on human facial feature information. A series of related technologies that use a video camera or camera to collect images or video streams containing human faces, automatically detect and track human faces in the images, and then perform facial recognition on the detected faces, usually also called portrait recognition and facial recognition. . [0003] The image recognition methods for target objects such as faces and objects in the prior art use a relatively simple rectangular feature template as the feature representation of the object. However, the rec...

AI Technical Summary

Problems solved by technology

[0003] The image recognition methods for target objects such as faces and objects in the prior art use a relatively simple rectangular feature template as the feature representation of the object. However, the recognition accuracy of this recognition method is not high; at the same time, the relatively simple rectangular The feature template also makes the final selection of representative rectangular feature templates more, resulting in more calculations required for classification

Therefore, the image recognition method in the prior art has the problems of low recognition accuracy and large amount of computation.

Images