Preserving Heritage Palaces: A Deep Learning CNN-SVM Hybrid Approach for Multi-classification

With the flow of time and enhancement in overall technology. People are disconnecting from their culture and the old palaces. This is very important for us to be connected with our old generations and learn the overall things from this from their architecture and experience. This all can be possible by studying the historical monuments and learning about our history. By increasing the saving historical palace moto this research can be done for predicting the different types of historical monuments with the help of machine learning and deep learning techniques and models. Convolution Neural Networks (CNN) may extract features using a variety of approaches without using specialized models, and over some time it can improve classification accuracy. To classify features from Heritage Palaces, a hybrid model was created that combines CNN and support vector machines (SVM). It is a double-layer Model. There image dataset of historical palaces has been collected from the web open source pics and created a dataset from it. As a result of web sources, a total of 4300 preprocessed images have been used for training and testing purposes in the CNN model. Many image preprocessing techniques have been implemented on it to make the dataset more functional. With the help of a combined CNN and SVM approach, a total of 97% classification accuracy is achieved by grand fountains, and Doric pillars feature classes. Traditional machine learning (ML) methods classify heritage palaces based on a few handmade items. That may result in an algorithm's execution failing, as well as reducing its efficiency. A glance revealed that the proposed hybrid model offers improved and more efficient strategies for classification.