Multiple Smile Detection Using Histogram of Oriented Gradient and Support Vector Machine Methods
Rezki Syaputra, Dedy Syamsuar, Edi Surya Negara · IOP Conference Series Materials Science and Engineering · 2021
Abstract The face or the front of the head consists of the eyes, nose and mouth. Each face has its uniqueness from the many human faces. The face is used to show happy and sad expressions and feelings. Smiling also includes self-expression from others. The analysis of facial expressions plays a key role in analyzing human emotions and behavior. Smile detection is a specialized task in facial expression analysis with a variety of potential applications such as photo selection, user experience analysis, smile payments, and patient monitoring. Conventional approaches often extract low-level face descriptors and smile detection based on strong binary classifiers. In this paper, we propose an effective Histogram of Oriented method supporting vector engines for smile detection. The experimental results show that our proposed network outperforms the current state of the method. The test results using the Histogram of Oriented Gradient and Support Vector Machine method for smile detection were 87% for a precision value of 88% and a recall value of 83% and accuracy. In the future, we want to exploit some of the latest effective designs. we will try to update our mouth model so we can support a bigger head turn and face size scale.