Detection method and device for objects in images

Abstract

The embodiment of the invention discloses a detection method and device for objects in images. The detection method and device are used for improving real timeliness of object detection. The method comprises the steps of inputting an image to be detected into a convolutional neural network which is previously trained, wherein the convolutional neural network is used for determining feature vectors corresponding to all candidate areas in the image; for the feature vectors corresponding to all the candidate areas, determining types of the objects contained in the candidate areas when the maximum values of type parameters in the feature vectors are larger than a set threshold value; determining position information of the objects contained in the candidate areas according to position parameters in the feature vectors. According to the detection method disclosed by the embodiment of the invention, the convolutional neutral network is utilized to determine the candidate areas in detection and calculate the feature vectors corresponding to all the candidate areas; thus, the type and the position of each object can be recognized, repeated operation is avoided, and real timeliness of detection is improved.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to an object detection method and device in an image. Background technique [0002] With the development of video surveillance technology, intelligent video surveillance is applied in more and more scenarios, such as traffic, shopping malls, hospitals, communities, parks, etc. The application of intelligent video surveillance is object detection through images in various scenarios Foundation. [0003] When performing object detection in an image in the prior art, a candidate region-based convolutional neural network (Region Convolutional Neural Network, R-CNN) and its extensions Fast RCNN and Faster RCNN are generally used. figure 1 It is a schematic diagram of the process of object detection using R-CNN. The detection process includes: receiving an input image, extracting a candidate region (region proposal) in the image, calculating the CNN feature of each candidate regi...

AI Technical Summary

Problems solved by technology

The above-mentioned detection of the category and position of the object is two independent steps. In both steps, the 256-dimensional data needs to be detected separately. Therefore, this process will also increase the detection time, thereby affecting the real-time performance of object detection.

Images