Large scale visual classification with SVM, create the unique article through NLP

摘要

Lack of Proper Information around Sri Lankan Historical Places cause to give false information. If ruins scattered around same place tourists are struggling to identify them. Sometimes there are small sign boards but they don't provide enough information. In this paper we address this challenge by using user captured image or search imaged. Our approach is produced a single document about identified places. System identify the places through training dataset using SVM. System will increase the accuracy of prediction by taking GPS data and user inputs necessarily. The evaluation shows that our approach is more accurate than the existing systems.