Haar Local Binary Pattern Feature for Fast Illumination Invariant Face Detection
read more
Citations
Local binary features for texture classification
Local Binary Patterns for Still Images
Feature representation for statistical-learning-based object detection
Research and Perspective on Local Binary Pattern
References
Statistical learning theory
Rapid object detection using a boosted cascade of simple features
Eigenfaces vs. Fisherfaces: recognition using class specific linear projection
A comparative study of texture measures with classification based on featured distributions
The FERET evaluation methodology for face-recognition algorithms
Related Papers (5)
Rapid object detection using a boosted cascade of simple features
Frequently Asked Questions (11)
Q2. What is the central concept of the Viola and Jones’ face detector?
The central concept of their framework (as in the Viola and Jones’ face detector) is to use boosting, that linearly combines simple weak classifiers f j(I) to build a strong ensemble, F(I) as follows :F(I) = n∑ j=1 α j f j(I).
Q3. how many false positives does HLBP have?
Although MCT is able to achieve an initial higher True Positive Rate (TPR), HLBP is able to outperform MCT as soon as the number of false positives are allowed to reach 50.
Q4. What is the main challenge for a face detection system?
The main challenge for a face detection system is to successfully detect faces in an arbitrary image, irrespective of variations in illumination conditions, background, pose, scale, expression and the identity of the person.
Q5. What are the advantages of the HLBP feature?
Their features are able to model the region-specific variations of local texture and are relatively robust to wide variations in illumination, pose and background, and also slight variations in pose.
Q6. How many times can the LBP4 image be calculated?
the calculation of the LBP4,1 image can be done in one pass over the image using only two relational operations per pixel.
Q7. What is the advantage of the Haar feature set?
Since each LBP label count is actually a particular bin value of the spatial histogram [24], their features are also robust to slight variations in location and pose.
Q8. What is the resulting decimal form of the LBP code?
The decimal form of the resulting 4-bit word is called the LBP code or LBP label of the center pixel and can be expressed as,ILBP(xc,yc) = 3∑ n=0 s(I(xn,yn)− I(xc,yc))2n.
Q9. What is the need to improve the system to take into account these illumination variations?
there is a need to improve the robustness of the system to take into account these illumination variations, but retaining the richness of the feature set, and the advantages of efficient feature selection by boosting and fast evaluation of the features using the Integral Image concept.
Q10. What is the final stage of the Integral Histogram set?
In the third and final stage, the Integral Histogram set will enable us to calculate the proposed HLBP features directly in an efficient and fast way as with Integral Image for the original Haar features.
Q11. What are the three types of HLBP features?
A particular HLBP feature is defined by the following parameters : mask type T (one out of five, ref. Fig. 3), LBP label k ( one out of sixteen for LBP4,1 ), position (x,y) of the mask inside the image plane, size (w,h) of the mask, a threshold θ and a direction p ( either +1 or -1 ).