Fighter Aircraft Detection using CNN and Transfer Learning
Motati Dinesh Reddy1, Sai Venkata Rao Kora2, Gnana Samhitha Ch.3
1Motati Dinesh Reddy, Department of Computer Science and Engineering, SRM Institute of Science and Technology, Kattankulathur (Tamil Nadu), India.
2Sai Venkata Rao Kora, Department of Electronics and Communication Engineering, SRM Institute of Science and Technology, Kattankulathur (Tamil Nadu), India.
3Gnana Samhitha Ch., Department of Computer Science and Engineering, SRM Institute of Technology, Amaravati (AP), India.
Manuscript received on 30 September 2022 | Revised Manuscript received on 10 October 2022 | Manuscript Accepted on 15 October 2022 | Manuscript published on 30 October 2022 | PP: 114-120 | Volume-12 Issue-1, October 2022 | Retrieval Number: 100.1/ijeat.A38541012122 | DOI: 10.35940/ijeat.A3854.1012122
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In this work, Deep learning techniques such as Convolutional Neural networks (CNN) and Transfer Learning are used to detect and identify Fighter aircraft or jets. A dataset consisting of 21 different aircraft with 20000 images is being processed using the above algorithms. CNN works on the principle of “pooling,” which progressively reduces the spatial size of the model to decrease the number of parameters and computations in the network. CNN’s are widely used for image detection in different domains, including defense, agriculture, business, face recognition technology, etc. Transfer learning is a machine learning method where a model created for a task is reused as the initial point for a model on a second task. Transfer learning is related to issues such as multi-task learning and concept drift and is not only an area of study in deep learning. The dataset is processed and uses python libraries such as pandas, seaborn, sci-kit- learn, etc., to find any pre-trained patterns and insights. Data is separated into train and test datasets with 80-20 percent of total data, respectively. A model is built using the TensorFlow library for CNN. The metric used is “accuracy.” A transfer learning model is also built to compare the accuracy results and adopt the best-fitting one.
Keywords: Convolution Neural Network, Deep Learning, Artificial Neural Network, Support Vector Machines
Scope of the Article: Deep Learning