Automatic Warship Recognition System : Dataset, Feature Representation and Classification Analysis
- Resource Type
- Conference
- Authors
- Kara, Yavuz Alper; Ucarer, Omer Kursat; Gundogdu, Batuhan
- Source
- 2019 27th Signal Processing and Communications Applications Conference (SIU) Signal Processing and Communications Applications Conference (SIU), 2019 27th. :1-4 Apr, 2019
- Subject
- Communication, Networking and Broadcast Technologies
Signal Processing and Analysis
Support vector machines
Marine vehicles
Dogs
Histograms
Optical imaging
Feature extraction
Principal component analysis
Warship Classification
Histogram of Oriented Gradients
Support Vector Machines
Feature Extraction
- Language
Classification of warships plays a critical role particularly in crises and war times. While there are several studies in the literature regarding classification of civilian ship types, warship classification task is yet far from maturity, which are significantly more similar to each other compared to civilian ships. In this study, we present a dataset and propose a system that employs automatic classification of warships based on their optical images. Histogram of Oriented Gradients (HOG) features extracted from ship images were investigated after several preprocessing steps which are then used in classification with Support Vector Machines (SVM). A dataset is composed based on images of particularly similar 9 warship classes that exist in the Turkish Navy and it has been shown that the proposed approach reaches 83:8% classification accuracy.