Panoramic street view privacy protection method based on aggregation channel features

Abstract

The invention provides a panoramic street view privacy protection method based on aggregation channel features, and relates to the field of computer vision. The method comprises the following steps that multichannel features are extracted, a classifier is trained and enhanced, feature combinations suitable for faces and license plates are obtained through experimental verification and detection of the faces and the license plates is unified; 10 channels in total including LUV, gradient magnitude and 6 gradient directions of histograms are obtained in the multichannel features through the experiment to act as a feature set of the optimal street view face detection and license plate detection computation speed and final classification accuracy; the channel features of scales near the scale are estimated according to the existing scale channel features; a target is self-adaptively blurred according to the size of the detection target; at least 2 weak classifiers are trained and finally a robust strong classifier is formed, and an improved rapid enhancing decision tree is used as the classifier to perform pruning in the early stage of training so that computation speed can be greatly enhanced; and the functions of background batch processing, click blurring and rapid deblurring are additionally arranged for plotting and other special application backgrounds.

Description

technical field

[0001] The invention relates to the field of computer vision, in particular to a privacy protection method for panoramic street views based on aggregated channel features. Background technique

[0002] In daily life, we often use maps when we travel. While using maps, we often view physical street view images. Usually, street view images are panoramic, but what we are presented here are generally ordinary images. In some special applications, such as surveying and mapping, it is also necessary to collect field pictures, including street view pictures. This kind of collection task usually uses a panoramic camera to collect panoramic pictures. Through the above pictures, we can obtain sufficient information, but at the same time, we do not need certain information or we do not want to obtain this information, such as pedestrian faces and vehicle license plates in the street view. At this time, we need to detect Face and license plate, blur the face and license...

Examples

Embodiment Construction

[0025] The following embodiments will further illustrate the present invention in conjunction with the accompanying drawings.

[0026] The embodiment of the present invention includes the following steps:

[0027] A. Extract multi-channel features to train the improved classifier, and the experiment verified that there is a feature combination suitable for both face detection and license plate detection, which unifies the framework of face and license plate detection;

[0028] At present, various artificial features have their advantages in some specific aspects. For example, SIFT features have rotation invariance and scale invariance, HOG (Dalal N, Triggs B. Histograms of oriented gradients for humandetection[C] / / Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on. IEEE, 2005, 1: 886-893) is a local statistical feature, suitable for the entire pedestrian detection. This paper extracts multiple channel features and uses multi-channel f...

technical field

[0001] The invention relates to the field of computer vision, in particular to a privacy protection method for panoramic street views based on aggregated channel features. Background technique

[0002] In daily life, we often use maps when we travel. While using maps, we often view physical street view images. Usually, street view images are panoramic, but what we are presented here are generally ordinary images. In some special applications, such as surveying and mapping, it is also necessary to collect field pictures, including street view pictures. This kind of collection task usually uses a panoramic camera to collect panoramic pictures. Through the above pictures, we can obtain sufficient information, but at the same time, we do not need certain information or we do not want to obtain this information, such as pedestrian faces and vehicle license plates in the street view. At this time, we need to detect Face and license plate, blur the face and license...

Examples

Embodiment Construction

[0025] The following embodiments will further illustrate the present invention in conjunction with the accompanying drawings.

[0026] The embodiment of the present invention includes the following steps:

[0027] A. Extract multi-channel features to train the improved classifier, and the experiment verified that there is a feature combination suitable for both face detection and license plate detection, which unifies the framework of face and license plate detection;

[0028] At present, various artificial features have their advantages in some specific aspects. For example, SIFT features have rotation invariance and scale invariance, HOG (Dalal N, Triggs B. Histograms of oriented gradients for humandetection[C] / / Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on. IEEE, 2005, 1: 886-893) is a local statistical feature, suitable for the entire pedestrian detection. This paper extracts multiple channel features and uses multi-channel f...

Images