Unlock instant, AI-driven research and patent intelligence for your innovation.

System and method of face recognition using portions of learned model

A technology for learning models and classification methods, applied in character and pattern recognition, image analysis, computer parts, etc., and can solve problems such as waste, waste of time and space

Inactive Publication Date: 2006-09-20
KONINK PHILIPS ELECTRONICS NV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

First, all effort related to training the rest of the network is wasted
Second, generalization performance on the validation set has a stochastic component due to noise in the data, so the network with the best performance on the validation set is not necessarily the one with the best performance on new or unseen test data
The obvious disadvantage of training multiple classifiers is that a lot of time and space are wasted in training and storing the model files

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • System and method of face recognition using portions of learned model
  • System and method of face recognition using portions of learned model
  • System and method of face recognition using portions of learned model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] For the sake of description, a radial basis function ("RBF") classifier is implemented, although any classification method / device may be implemented. The RBF classifier device is a co-owned, co-pending U.S. Patent Application Serial No. filed on February 27, 2001, entitled "Classification of objects through model ensembels". 09 / 794,443 (published as WO02 / 069267) is available, and the entire content and disclosure of the patent application are hereby incorporated by reference in their entirety.

[0017] Now refer to figure 1 Describes the structure of the RBF network disclosed in the co-owned, co-pending U.S. Patent Application Serial No. 09 / 794,443. Such as figure 1 As shown, the basic RBF network classifier 10 is constructed according to a traditional three-layer backward propagation network, including a first input layer 12 composed of source nodes (for example, k sensing units); including i nodes (its function is to gather The second or hidden layer 14 of class data an...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A system and method for classifying facial image data, the method comprising the steps of: training a classifier device for recognizing one or more facial images and obtaining corresponding learned models the facial images used for training; inputting a vector including data representing a portion of an unknown facial image to be recognized into the classifier; classifying the portion of the unknown facial image according to a classification method; repeating inputting and classifying steps using a different portion of the unknown facial image at each iteration; and, identifying a single class result from the different portions input to the classifier.

Description

Technical field [0001] The present invention relates to a facial recognition system, in particular, to a method of performing facial recognition by using a part of a learning model. Background technique [0002] Existing facial recognition systems try to recognize the unknown face by matching the unknown face with previous instances of that subject's face. This is typically by training a classifier using previous instances of the subject's face, and then using the trained classifier to identify the subject by matching the previous instance of the subject's face with a new instance of the subject's face. As we all know, training a classifier involves learning a model of the subject's face. Existing systems use the entire model during classification. [0003] Although the ultimate goal of any pattern recognition system design is to obtain the best possible classification (prediction) performance, this goal has traditionally led to the development of different classification schemes...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06T7/00
CPCG06K9/00288G06V40/172G06V10/774
Inventor S·V·R·古特塔V·菲尔洛明M·特拉科维
Owner KONINK PHILIPS ELECTRONICS NV