To search, Click below search items.

 

All Published Papers Search Service

Title

A New Formulation of Face Sketch Multiple Features Detection Using Pyramid Parameter Model and Simultaneously Landmark Movement

Author

Arif Muntasa, Mochamad Hariadi, Mauridhi Hery Purnomo

Citation

Vol. 9  No. 9  pp. 249-260

Abstract

There are many cases when crimes happened, the criminals can not be identified. But with the help of the eyewitness reconstruction, the sketchers can make the suspects’ face to help the identification. With this in mind, we propose a new formulation of face sketch multiple features detection. It is conducted by using the Pyramid Parameter Model and simultaneously landmark movements’ improvement. Our proposed method is divided into four stages: training, create image gradient, shape initialization, and multiple features detection processes. The last stage is started by searching the maximum line gradient value between two landmarks. Thus, by using the Similarity Transformation Equation, the set of landmarks (shape) will be simultaneously moved. The movements are conducted by reducing the range and parameter value on each movement. This process is called Pyramid Parameter Model. The position of new landmark is enhanced by using simultaneously landmark movements on each shape. As the testing set, we use 50 halftone and 50 hatching face sketches. Each sketch is examined by using 38 landmarks of 7 features. The experimental results show that the average of detection rate is 93.28% for halftone and 93.03% for hatching face sketch.

Keywords

Pyramid Parameter Model, Maximum Line Gradient, Multiple Features Detection, Halftone and Hatching Face Sketches, police investigation

URL

http://paper.ijcsns.org/07_book/200909/20090931.pdf