00000ctmaa22000003a 4500 UP-1685675941123958178 Buklod 20180810111032.0 aa rb |||1 u| ta 180810s xx d r |||| u| (iLib)UPTAC-00019010296 UPVTC eng LG 993.5 2018 C66 C33 Cabahug, Warren Dave S. author. Detectibee digital image splicing detection using artificial bee colony-trained artificial neural networks Warren Dave S. Cabahug; Ryan Rey M. Daga, adviser. 2018. ix, 102 leaves illustrations. Undergraduate thesis (B.S. Computer Science) -- University of the Philippines, Tacloban. The rapid growth and ease of access to powerful image editing tool is one of the major factors that led to the increase in number of altered and manipulated images circulated. To determine the authenticity of images, automatic forgery detection algorithms have been developed. One of the most common digital image forgery is digital image splicing which can be easily performed by compositing two or more images. In this study, an image splicing forgery detection scheme was presented. Chrominance components were introduced in this work as compared to the commonly used RGB and luminance spaces. Features extracted from the gray-level co-occurrence matrices (GLCM) of the decor related chrominance components were used to train the classifier. The Artificial Bee Colony (ABC) algorithm, which is a recent swarm-based numerical optimization algorithm inspired buy the intelligent foraging behavior of the honeybees was used to train the Artificial Neural Network (ANN) which was employed as a classifier to demonstrate the performance of the proposed scheme. The accuracy rate was used to evaluate the effectiveness of the proposed scheme. Experimental results have shown that the GLCM features extracted from the chroma channels provide much better performance than those from the RGBB and luminance channels. Furthermore, the ABC-trained ANN performed well in discriminating authentic and forged images. Digital image splicing. Artificial bee colony algorithm. Artificial neural network. Daga, Ryan Rey M. adviser. FI UP UPTAC UPTAC LG 993.5 2018 C66 C33 Thesis