PERANCANGAN SISTEM BIOMETRIK TELINGA BERBASIS GRAY LEVEL CO-OCCURRENCE MATRIX MENGGUNAKAN JARINGAN SYARAF TIRUAN PERAMBATAN BALIK

In Indonesia, cases such as accidents, fires, and criminal acts frequently result in loss of life. Forensic science is essential as a process for identifying both living and deceased individuals, to facilitate the investigation of victims and perpetrators of crimes. In certain situations, where a corpse is found in a dismembered or mutilated condition, the shape, size, and individual characteristics of the ears are useful in identifying the victim along with other identifying features of the human body. Among the various biometric fields, the ear is the most unique part of a person's body, with parameters such as contour, appearance, and posture. In this study, a total of 130 images were taken from primary data using a Redmi 9 Plus camera, which captured 48 images, and a Xiaomi Note 9 Pro, which captured 52 images, from each of 5 individuals, totaling 16 images. For training data, 80 images were used, while 50 images were used for testing data. Feature extraction using Gray Level Co-occurrence Matrix. (GLCM). The identification method using backpropagation Neural Networks achieved an accuracy of 90%.%.

Keywords : Backpropagation Artificial Neural Network , Biometrics, Gray Level Co-Occurrence Matrix, , Image Processing, Ear