Design method for multi-pose human face detector based MSNRD feature

Abstract

The invention discloses a design method for a multi-pose human face detector based MSNRD features. The design method comprises a training stage and a detection stage. The training stage comprises the steps of: determining size of a template at first, and preparing a training data set; then constructing a featurepool; and training a strong classifier under an Adaboost framework. The detection stage comprises the steps of: traversing subimages of each scale at each position of an input image at first, judging whether the image is a human face; forming a set composed of all human face images, reserving elements with large output values of the classifier if a ratio of an overlapping area of any two elements in the set to a minimal element area exceeds 0.3, repeating the step till the ratio of the overlapping area of any two elements in the set to the area of a smaller one does not exceed 0.3, and regarding the set as a final result of detection for output. The design method provided by the invention enriches the feature expression capability, simplifies the feature calculation process, avoids the training of detectors respectively for different poses, reduces training workload, and increases the detection efficiency.

Description

technical field [0001] The invention relates to a design method of a multi-attitude human face detector, in particular to a design method of a multi-attitude human face detector based on MSNRD features. Background technique [0002] The face detector is the most basic step in the face recognition process, and its performance is directly related to the recognition accuracy of the recognition system. The main factors restricting its performance are: face posture, expression, and the richness and diversity of the environment. At present, the mainstream method of face detector is Haar or LBP features combined with Adaboost to form a strong classifier, and then multiple strong classifiers are cascaded to form a waterfall structure. LBP’s resistance to noise is weak, resulting in unstable detection results, which is not conducive to subsequent processing of faces; Haar features have two obvious defects: first, the expression ability is limited, limited to the expression of points...

AI Technical Summary

Problems solved by technology

LBP’s resistance to noise is weak, resulting in unstable detection results, which is not conducive to subsequent processing of faces; Haar features have two obvious defects: first, the expression ability is limited, limited to the expression of points, lines and edges, The face patterns are rich and diverse, so the expressive ability of the features is required to be strong; second, the calculation of the Haar feature requires integral graphs, square integral graphs, two additions, two divisions, and two root operations, which require a large amount of calculation and greatly Affected the detection speed

[0003] In order to improve the adaptability of the detector to face pose changes, expression changes and ambient light, the existing public technologies are to build different detectors for different poses, and the final result is the union of the detection results of multiple detectors; this exists Three flaws: First, faces are rich in variations, and it is difficult to classify and construct datasets, and the workload of classifying and constructing datasets is huge; second, face images with different poses, expressions, and lighting conditions have some common features. If the training is divided into categories, it is not conducive to make full use of these common features to improve the detection efficiency; third, since the detector scans one by one in different positions and different scales of the image in the form of a sliding window in actual use, if there are too many detectors , will inevitably affect the detection speed, which is not conducive to the application of the detector in real-time occasions

Images