Abstract
Casting is one of the most important processes applied within various kinds of industries. The strength of casted products could be affected dramatically if the casting underwent any defects. Hence, these defects must be detected and resolved. The detection is usually carried out through the inspection of the X-ray casting images. Existing methods examine the general distribution of defects by extracting and classifying features from the casting X-Ray image. In this paper, distortion of the X-ray casting image is considered and an efficient feature extraction algorithm is evaluated based on the Cepstral Coefficients technique. The algorithm uses coefficients from the Discrete Cosine Transform (DCT), Discrete Sine Transform (DST), and Discrete Wavelet Transform (DWT). Furthermore, the Artificial Neural Network (ANN) is employed for the classification phase. In this work, the features are extracted from X-ray image and its transformed versions. Performance comparisons are applied between different transforms in terms of their impact on the achieved recognition rate. The experimental results show that extracting features using DCT offers a higher recognition rate compared to other transform domains.