Limits...
A high precision feature based on LBP and Gabor theory for face recognition.

Xia W, Yin S, Ouyang P - Sensors (Basel) (2013)

Bottom Line: A maximum improvement of 29.41% is achieved comparing with other methods.Besides, the ROC curve provides a satisfactory figure.Those experimental results strongly demonstrate the feasibility and superiority of the new feature and method.

View Article: PubMed Central - PubMed

Affiliation: Tsinghua Center for Mobile Computing, Institute of Microelectronics, Tsinghua University, Beijing 100084, China. maxiaola@gmail.com

ABSTRACT
How to describe an image accurately with the most useful information but at the same time the least useless information is a basic problem in the recognition field. In this paper, a novel and high precision feature called BG2D2LRP is proposed, accompanied with a corresponding face recognition system. The feature contains both static texture differences and dynamic contour trends. It is based on Gabor and LBP theory, operated by various kinds of transformations such as block, second derivative, direct orientation, layer and finally fusion in a particular way. Seven well-known face databases such as FRGC, AR, FERET and so on are used to evaluate the veracity and robustness of the proposed feature. A maximum improvement of 29.41% is achieved comparing with other methods. Besides, the ROC curve provides a satisfactory figure. Those experimental results strongly demonstrate the feasibility and superiority of the new feature and method.

Show MeSH

Related in: MedlinePlus

Databases. The six databases are, from top to bottom in turn: ORL, YALE, ABERDEEN, AR, FSTAR, FRGC, FERET (Reprinted with permission).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3673096&req=5

f7-sensors-13-04499: Databases. The six databases are, from top to bottom in turn: ORL, YALE, ABERDEEN, AR, FSTAR, FRGC, FERET (Reprinted with permission).

Mentions: In the recognition module, we choose six well lnown databases and a self-made database to test our new feature. Each has its different emphasis, as shown in Figure 7. The ORL Database consists of 400 images of 40 different persons. The images mainly vary in pose and scale. The Yale Database contains 15 individuals with 11 images of each one. It emphasizes illumination and expression changes. There are 3,288 images corresponding to 116 people's faces in the AR Database which shows dramatic changes of lighting, expressions, poses and even accessories such as glasses and bangs. We also randomly choose three images of each person as training set and the rest as testing set. The ABERDEEN database includes 625 images of 65 persons. The images are strictly controlled in different lighting and expression conditions. The FRGC Database contains 12,776 images in the training set and 8,014 images in the testing set. This database is quite a challenge for its severe variation. The last database is FERET, which includes 1,400 cropped gray images of 200 persons. Angles and expressions are the focus. It should be noted that in our daily life, rather than the natural factors such as light or poses, makeup is the one that mostly distracting recognition problems. Noticing this fact, we downloaded images of film stars with different makeup from the internet and gather them together to form a database named FSTAR database. The FSTAR database contains 1,800 images of 120 persons. For each database, a random subset with three images per individual is used for training and the rest for testing. When the subset changed, we regard it as a new group of experiments. We performed thousands of groups of experiments with random subsets.


A high precision feature based on LBP and Gabor theory for face recognition.

Xia W, Yin S, Ouyang P - Sensors (Basel) (2013)

Databases. The six databases are, from top to bottom in turn: ORL, YALE, ABERDEEN, AR, FSTAR, FRGC, FERET (Reprinted with permission).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3673096&req=5

f7-sensors-13-04499: Databases. The six databases are, from top to bottom in turn: ORL, YALE, ABERDEEN, AR, FSTAR, FRGC, FERET (Reprinted with permission).
Mentions: In the recognition module, we choose six well lnown databases and a self-made database to test our new feature. Each has its different emphasis, as shown in Figure 7. The ORL Database consists of 400 images of 40 different persons. The images mainly vary in pose and scale. The Yale Database contains 15 individuals with 11 images of each one. It emphasizes illumination and expression changes. There are 3,288 images corresponding to 116 people's faces in the AR Database which shows dramatic changes of lighting, expressions, poses and even accessories such as glasses and bangs. We also randomly choose three images of each person as training set and the rest as testing set. The ABERDEEN database includes 625 images of 65 persons. The images are strictly controlled in different lighting and expression conditions. The FRGC Database contains 12,776 images in the training set and 8,014 images in the testing set. This database is quite a challenge for its severe variation. The last database is FERET, which includes 1,400 cropped gray images of 200 persons. Angles and expressions are the focus. It should be noted that in our daily life, rather than the natural factors such as light or poses, makeup is the one that mostly distracting recognition problems. Noticing this fact, we downloaded images of film stars with different makeup from the internet and gather them together to form a database named FSTAR database. The FSTAR database contains 1,800 images of 120 persons. For each database, a random subset with three images per individual is used for training and the rest for testing. When the subset changed, we regard it as a new group of experiments. We performed thousands of groups of experiments with random subsets.

Bottom Line: A maximum improvement of 29.41% is achieved comparing with other methods.Besides, the ROC curve provides a satisfactory figure.Those experimental results strongly demonstrate the feasibility and superiority of the new feature and method.

View Article: PubMed Central - PubMed

Affiliation: Tsinghua Center for Mobile Computing, Institute of Microelectronics, Tsinghua University, Beijing 100084, China. maxiaola@gmail.com

ABSTRACT
How to describe an image accurately with the most useful information but at the same time the least useless information is a basic problem in the recognition field. In this paper, a novel and high precision feature called BG2D2LRP is proposed, accompanied with a corresponding face recognition system. The feature contains both static texture differences and dynamic contour trends. It is based on Gabor and LBP theory, operated by various kinds of transformations such as block, second derivative, direct orientation, layer and finally fusion in a particular way. Seven well-known face databases such as FRGC, AR, FERET and so on are used to evaluate the veracity and robustness of the proposed feature. A maximum improvement of 29.41% is achieved comparing with other methods. Besides, the ROC curve provides a satisfactory figure. Those experimental results strongly demonstrate the feasibility and superiority of the new feature and method.

Show MeSH
Related in: MedlinePlus