Klasifikasi Citra Tekstur Daging Sapi, Kambing, dan Babi Menggunakan Ekstraksi Fitur Wavelet Haar dan Symlet Berbasis Support Vector Machine

Green Kenny Sarimanella (UNIVERSITAS PATTIMURA)
Francis Yunito Rumlawang (Universitas Pattimura)
Harmanus Batkunde (Universitas Pattimura)
Meilin Imelda Tilukay
A. Z. Wattimena

Tensor: Pure and Applied Mathematics Journal
Vol 7 No 1 (2026): Tensor: Pure and Applied Mathematics Journal

Meat is one of the animal protein sources widely consumed by the public; however, distinguishing different types of meat visually is often difficult because they have very similar textures. This study applies the Support Vector Machine (SVM) method with feature extraction based on Haar Wavelet and Symlet Wavelet (Sym4) to classify texture images of beef, goat meat, and pork. The dataset consisted of 1200 digital images processed through resizing, grayscale conversion, and normalization stages. Feature extraction was performed using the Discrete Wavelet Transform (DWT) to obtain statistical texture features. The classification process employed the Radial Basis Function (RBF) kernel with a multiclass classification approach. The results showed that the Haar Wavelet achieved an accuracy of 96.67%, while the Symlet Wavelet (Sym4) achieved 94.17%. These findings indicate that the combination of wavelet methods and SVM is effective for automatic and objective meat type identification