Abstract: In this work we present an algorithm designed to recognize if the faces in a still image are smiling or not, based on Viola Jones face detection method and an algortihm for smile recognition. The stages of our algorithm are: identification of the faces in the image, detection of the mouth part of each face and then application of a basic algorithm to the mouth part to determine whether the person is smiling or not. A smile is recognized by examining the graph of the grayscale levels across the mouth and looking for two distinct minima. As an additional step of image quality assessment, we measure the level of blurriness of the detected face. Real life images were used for testing the algorithms at different spatial resolutions, levels of blurriness as well as varying contrast and brightness conditions. Our conclusion is that the proposed methods can be beneficial for automatic smile recognition and blur detection in digital cameras or other mobile devices.

• Publication Date: 04-Aug-2013
• DOI: 10.15224/978-981-07-7227-7-06
• Views: 0
• Downloads: 0

Abstract: Grammatical Evolution (GE), which is one of the evolutionary computations, aims to find function, program or program segment satisfying the design objective. This paper describes the improvement of the Grammatical Evolution according to Stochastic Schemata Exploiter (GE-SSE) and its application to symbolic regression problem. Firstly, GE-SSE is compared with original GE in symbolic regression problem. The resultsshow that GE-SSE has faster convergence property than original GE. Secondly, GE-SSE is applied for the stock price prediction as the actual application of the GE-SSE.

• Publication Date: 04-Aug-2013
• DOI: 10.15224/978-981-07-7227-7-08
• Views: 0
• Downloads: 0