Image object detection method

Abstract

The invention provides an image object detection method. The image object detection method is capable of achieving good detection performance in a large image data set. The method includes the steps that multiple sample images are marked according to the priorities of information amounts, so that corresponding marked images are obtained; regions containing objects or regions highest in probability of object existence are extracted from the marked images, and candidate windows are generated; feature expressions of the candidate windows are extracted from a convolution neural network and form a candidate set, the candidate set is fitted through semi-supervised learning, and therefore a target function of an image detection model is obtained; the region containing a target object or the region highest in probability of target object existence is extracted from an image to be detected, a window to be detected is generated, the feature expression of the window to be detected is extracted and detected, and the candidate window highest in the probability of target object existence is obtained. The method can rapidly and accurately detect the target object in the large data set.

Description

technical field [0001] The invention relates to the technical field of image recognition or image processing, in particular to an image object detection method. Background technique [0002] In image recognition or image processing technology, image object detection is widely used, such as crime tracking, crowd counting and analysis of large-scale sports events or expositions, smart cities, smart transportation, smart homes, online shopping retrieval, image search, image or video Semantic real-time understanding, etc. Whether it can better recognize input images such as faces and find matching images with high correlation, on the one hand, depends on whether the image database is complete enough. Massive image and video data will help improve retrieval performance; on the other hand, with It depends on whether the image detection method used is appropriate, which requires the computer to model the image data set reasonably, so that the most similar search results can be qui...

AI Technical Summary

Problems solved by technology

[0004] The unsupervised learning method is based on a data set without prior information labels, which is not conducive to obtaining faster calculation speed under limited computing power, and it is difficult to obtain good detection performance; the convolutional neural supervised learning method is based on complete prior information labels. The implementation of the dataset helps to improve the performance of image detection, but it is not suitable for image datasets with large sample size due to the limitation of labeling labor costs and hardware storage capacity.

[0005] The semi-supervised learning method is based on a data set with some images attached to the label, which consumes moderate personnel and hardware resources. However, with the increase of image data, the fuzziness of the data set in the existing image detection model increases. The target obtained through semi-supervised learning The fit of the function to large data sets is reduced, so the existing semi-supervised learning image object detection methods have poor performance in detecting images on large image data sets

Images