Detecting Fighter Jets using Convolutional Neural Networks and TransferLearning

Deep learning methods such as Convolutional Neural Networks (CNN) and Transfer Learning are utilized in this research to locate and recognize fighter jets and aircraft’s. The aforementioned methods are currently being applied to a dateset that contains twenty distinct airplanes and mine thousand plus photographs. The ”pooling” idea, Which gradually reduces the spatial dimension of the model in order to decrease the number of parameters and computations in the network, is the foundation upon which CNN is built. CNN’s are utilized extensively for the purpose of image detection across a variety of industries, including the military, agriculture, business, and other forms of face recognition technology. Transfer learning is a method of machine learning in which a model that was developed for one job is reused as the starting point for a model that is being developed for a different task. Transfer learning is not only a topic of research in the field of deep learning; it is also connected to problems such as learning several tasks at once and concept drift. The information is processed, and various Python modules, including pandas, seaborn, sci-kit learn, and others, are utilized in an effort to locate any per-trained patterns and insights. The data is divided into a train datasets and a test datasets, each of which contains 80 and 20 percent of the total data. The TensorFlow library for CNN is utilized in the construction of a model. ”Accuracy” is the metric that is employed. Additionally, a transfer learning model is constructed in order to evaluate the various accuracy results and select the most appropriate one.