TRANSFER LEARNING BASED CLASSIFICATION OF SEGMENTED LANDING PAGE COMPONENTS

Authors : Çağla ŞENEL, Gülşah AYHAN, Zeynep Eda URAN, Behçet Uğur TÖREYİN

Pages : 36-39

View : 15 | Download : 9

Publication Date : 2018-12-31

Article Type : Research Paper

Abstract :The pages that appear in front of users on digital platforms used for online advertising to attract attention to target product are called landing pages. Landing pages aim to increase advertisement conversion rate using the metrics like clicks, views or subscribes. In this study, a method is presented to automatically classifier the most commonly used components on landing pages which are buttons, texts, and checkboxes. Landing page images given as inputs are segmented by morphological and threshold-based image processing methods, and each segment is classified using a Transfer Learning based method which combines pre-trained Inception v-3 networks and Support Vector Classifier insert ignore into journalissuearticles values(SVM);. Furthermore, different classifiers were applied to compare the results. The proposed method is anticipated to be an essential step in the process of designing landing pages automatically with high advertisement conversion rates. Thanks to the proposed transfer learning based method, this is achieved by using fewer number of training data.
Keywords : Landing Page Segmentation, Transfer Learning, Image Processing, Image Classification