Facial Expression Recognition Using Diagonal Crisscross Local Binary Pattern

Main Article Content

T. Sabhanayagam, Dr. V. Prasanna Venkatesan, Dr. K. Senthamaraikannan

Abstract

Facial expression analysis is a noteworthy and challenging problem in the field of Computer Vision, Human-Computer Interaction and Image Analysis. For accomplishing FER, it is very difficult to acquire an effective facial description of the original facial images. The Local Binary Pattern (LBP) which captures facial attributes locally from the images is broadly used for facial expression recognition. But conventional LBP has some limitations. To overcome the limitations, novel approach for Facial Expression Recognition based Diagonal Crisscross Local Binary Pattern (DCLBP). It is based on the idea that pixel variations in diagonal as well as vertical and horizontal (crisscross) should be taken as an image feature in the neighborhood different from the other conventional approaches.The Chi-square distance method is used to classify various expressions. To enhance the recognition rate and reduce the classification time, weighted mask is employed to label the particular components in the face like eyebrow, mouth and eye with larger weights than the other parts of the face. The results of comparison showed the performance of the suggested approach comparing to the other approaches and the experimental results on the databases JAFFE and CK exhibited the better recognition rate.

Article Details

How to Cite
, T. S. D. V. P. V. D. K. S. (2018). Facial Expression Recognition Using Diagonal Crisscross Local Binary Pattern. International Journal on Future Revolution in Computer Science &Amp; Communication Engineering, 4(3), 376–382. Retrieved from http://www.ijfrcsce.org/index.php/ijfrcsce/article/view/1325
Section
Articles