<![CDATA[Humans can quickly and make accurate predictions from visual images. Among facial tasks, gender classification is one that has an important role and is probably the easiest and fastest way to achieve. This time, the use of computers continues to grow so that a system similar to human capabilities is built, namely the gender recognition based on facial images. Some applications that require a gender recognition system such as, application of human-computer interaction interface (adjusting software behavior concerning the gender of the user), and demographic data collection to determine trends and product recommendations in the store based on gender. The use of accessories on facial images such as glasses, earrings, and hats that can make a person's gender difficult to recognize is a challenge in doing gender classification based facial image by the system. Compass Local Binary Patterns (CoLBP) as one of the image processing methods used in feature extraction for gender classification based face images. CoLBP utilizes the Kirsch Compass Mask to improve the performance of Local Binary Patterns (LBP) in the feature extraction process. In this research using the Color FERET dataset containing photos of faces (with accessories and without accessories) and the Random Forest classification method for the evaluation process. In the test results, the best accuracy average is 91.8%. From this research, can be concluded that the CoLBP method provides good feature extraction performance and accessories on the face give an influence on the reducing quality of the feature extraction by the CoLBP method.]]>
Copyrights © 2019
