Image Retrieval Based on Self-Organizing Feature Map and Multilayer Perceptron Neural Networks Classifier
- Resource Type
- Conference
- Authors
- Ghaleb, Moshira S.; Ebied, Hala M.; Shedeed, Howida A.; Tolba, Mohamed F.
- Source
- 2019 Ninth International Conference on Intelligent Computing and Information Systems (ICICIS) Intelligent Computing and Information Systems (ICICIS), 2019 Ninth International Conference on. :189-193 Dec, 2019
- Subject
- Bioengineering
Communication, Networking and Broadcast Technologies
Computing and Processing
Robotics and Control Systems
Signal Processing and Analysis
Feature extraction
Self-organizing feature maps
Neurons
Image color analysis
Image retrieval
content-based image retrieval
feature extraction
neural networks
self-organized map
multilayer perceptron
- Language
Content-based image retrieval (CBIR), is a type of computer vision application. It searches for an image by image content not text in large database images. According to the huge existence of image databases, the searching time and high accuracy retrieved images become a great challenge. This paper aims to find a solution to retrieve images with high accuracy results. Neural network became a hot topic in the image processing field for the past few years. This paper presents two approaches to Content-based image retrieval. The first approach used the Self-Organized Feature Map (SOFM) as a clustering method to image retrieval. The second approach consists of two phases. The first one used the SOFM as a feature extraction method. The second phase used the Multilayer Perceptron (MLP) as a classifier method. The paper studied the impact of changing some parameters values on recognition accuracy. The experiments carried out using the Wang Corel 1000 database. The results show that the SOFM+MLP improved the recognition accuracy compared to SOFM. The SOFM+MLP achieved approximately 99% average recognition accuracy.