Identification of Flower Species with Natural Markers
Previous studies have shown that improvements in image description methods have centred on the continuous interest in retrieving the best relevant images to the query image. This has been done using various specimens of images across different disciplines. The aim of this study in order to study and...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Universiti Malaysia Sarawak (UNIMAS)
2019
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/27456/1/Tan%20Mei.pdf http://ir.unimas.my/id/eprint/27456/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Malaysia Sarawak |
| Language: | English |
| Summary: | Previous studies have shown that improvements in image description methods have centred on the continuous interest in retrieving the best relevant images to the query image. This has been done using various specimens of images across different disciplines. The aim of this study in order to study and analyst the unique features of flower species and to propose a window-based selection technique for improvement the efficiency in flower species identification. Algorithms of investigation span from extraction, matching and identification to determine the image features. The current study used flower images as a case study to illustrate the effectiveness of the proposed framework. The accurate identification of image features for flower species can reduce the complexity of processing, processing time and enhance the accuracy of an identification. The proposed method can successfully reduce the index value of features by 51.6%. In addition, the complexity of processing can be reduced from O(n log n) to O(n) after window selection phase, and the percentage of average accuracy for identifying flower species reached 98.8%. Hence, the total number of loops decreased from 1,500 times to 500 times in the CBIR system due to its faster output. |
|---|