المفاهيم الأساسية
Utilizing deep learning for automatic location detection through image classification.
الملخص
The article delves into the implementation of an image classification system using deep learning to automatically detect and classify images of Indian cities. The study focuses on classifying images of Ahmedabad, Delhi, Kerala, Kolkata, and Mumbai based on distinct features. Two approaches were adopted - a vanilla Convolutional Neural Network (CNN) and transfer learning with the VGG16 model. The VGG16 model achieved a test accuracy of 63.6%, showcasing the potential for real-time location identification systems. The research aims to contribute to tourism, urban planning, and other applications by recognizing city imagery's unique characteristics.
الإحصائيات
The VGG16 model achieved a test accuracy of 63.6%.
Dataset split: training set (70%), validation set (15%), test set (15%).
اقتباسات
"Our findings demonstrate the potential applications in tourism, urban planning, and even real-time location identification systems."
"The jump in accuracy from the Vanilla CNN to the VGG16 model emphasized the benefits of leveraging pre-trained architectures."