Real-time face recognition method based on deep neural network
A deep neural network and face recognition technology, applied in the face recognition field, which can perform face recognition tasks in real time, and can solve the problem of increasing the difficulty of deep network models, large image size, and poor applicability of face recognition, etc. question
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2014-05-07
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the field of pattern recognition, relates to a face recognition method, in particular to a face recognition method capable of performing face recognition tasks in real time. Background technique
[0002] As a kind of biometric authentication technology, face recognition has huge market potential and scientific research value due to its characteristics of non-contact, good user experience, and steadily rising recognition rate. Face recognition is a kind of image recognition. The most important and difficult point of image recognition is to give the machine the ability to understand the hidden information contained in the image. As a feature extraction method that can extract deep information from data, the deep neural network can be used for image-based human Face recognition technology has some inspiration.
[0003] At present, the deep neural network has made many breakthroughs in the field of pattern recognition: Microsoft u...
Examples
Embodiment Construction
[0058] The embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings: this embodiment is implemented on the premise of the technical solution of the present invention, and provides detailed implementation methods and specific operating procedures.
[0059] like figure 1 As shown, this embodiment includes the following steps:
[0060] Step 1. Obtain network training data, specifically: use the LFW face database with better diversity as the network training database in the unsupervised process, see figure 2 , using some images in CMU-PIE, Georgia Tech, CaltechFaces and VidTiMIT face database to combine into a mixed face database (containing 2311 images, each individual contains multiple images in multiple states), see image 3 , as the training data in the supervised process, in which the illumination normalization operation is performed on some images with strong illumination changes to reduce the influence of illumin...