Content based image retrieval for multi-objects fruits recognition using k-means and k-nearest neighbor
| Main Author: | Erwin, Erwin |
|---|---|
| Format: | Proceeding PeerReviewed Book |
| Bahasa: | eng |
| Terbitan: |
, 2017
|
| Subjects: | |
| Online Access: |
http://repository.unsri.ac.id/29087/1/Proceeding-erwin.pdf http://repository.unsri.ac.id/29087/2/IEEE%20Xplore%20Search%20Results.pdf http://repository.unsri.ac.id/29087/ |
Daftar Isi:
- The uniqueness of fruits can be observed using the colors and shapes. The fruit recognition process consists of 3 stages, namely feature extraction, clustering, and recognition. Each of stage uses different methods. The color extraction process using Fuzzy Color Histogram (FCH) method and shaping extraction using Moment Invariants (MI) method. The clustering process uses the K-Means Clustering Algorithm. The recognition process uses the k-NN method. The Content-Based Image Retrieval (CBIR) process uses image features (visual contents) to perform image searches from the database. Experimental results and analysis of fruit recognition system obtained an accuracy of 92.5% for single-object images and 90% for the multi-object image.