Target detection method and apparatus, and mobile terminal

Abstract

The invention, which belongs to the technical field of image classification detection, discloses a target detection method and apparatus, and a mobile terminal. The method comprises: a plurality of training samples are inputted and training is carried out repeatedly to obtain a classification parameter of a cascade classification model; according to the classification parameter, screening is carried out on an image needing detection to obtain a plurality of sorted candidate windows; training is carried out to generate a corresponding convolutional neural network model and then training is carried out by using the convolutional neural network model to form a corresponding classifier; and according to the classifier after training, all sorted candidate windows are detected to obtain an output result of target detection. In addition, the apparatus consists of a firs training unit, a screening unit, a second training unit, and a detection unit. According to the technical scheme, the method and apparatus have the following beneficial effects: a data calculation load of window obtaining is reduced and the detection efficiency of window obtaining is improved, so that the target detection algorithm becomes simple and efficient. Moreover, generalization of targets on which target detection is carried out is enhanced, so that target detection is suitable for a mobile terminal.

Description

technical field [0001] The present invention relates to the technical field of image classification and detection, in particular to a target detection method and device and a mobile terminal. Background technique [0002] Object detection is an important technology involving computer vision and image processing. Its main purpose is to detect object instances of certain categories (such as humans, buildings, vehicles, etc.) in images and videos. Object detection technology is currently used in image retrieval, video surveillance, and human-computer interaction. [0003] Existing target detection technologies usually have relatively large limitations, and due to the use of sliding window detection methods, the amount of data to be detected and calculated is large, resulting in increased computational complexity and reduced detection efficiency. Contents of the invention [0004] According to the problems existing in the prior art, a technical solution of a target detection ...

AI Technical Summary

Problems solved by technology

[0003] Existing target detection technologies usually have relatively large limitations, and due to the use of sliding window detection methods, the amount of data that needs to be detected and calculated is large, resulting in increased computational complexity and reduced detection efficiency

Images