Location-based services for mobile devices
The explosive growth of mobile technology is widely used for various opportunities in mobile media application development. Mobile apps which involve content-based visual information retrieval allow users to retrieve the information immediately through images. This project aims to transform a mobile...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/64672 |
| _version_ | 1811696124981936128 |
|---|---|
| author | Thin, Mya Mya |
| author2 | Yap Kim Hui |
| author_facet | Yap Kim Hui Thin, Mya Mya |
| author_sort | Thin, Mya Mya |
| collection | NTU |
| description | The explosive growth of mobile technology is widely used for various opportunities in mobile media application development. Mobile apps which involve content-based visual information retrieval allow users to retrieve the information immediately through images. This project aims to transform a mobile device into a portable information terminal that renders image recognition services for mobile devices. This project is to explore and evaluate various image recognition techniques to achieve high performance accuracy with shortest cost (execution time) on stamp image by experimenting the existing image recognition tools. This final year project is for mobile users to be able to identify different stamps of various countries by capturing the images of unknown stamps. In this project, the area of focus is to create a visual stamp database consisting of reference and test databases to perform image recognition process by using the Matlab program. Image recognition tools adopted the Bag-of-Words system and the Scale-Invariant Feature Transform (SIFT) is used for feature extraction. And also, the scalable vocabulary tree and hierarchical k-means are used for machine learning and clustering accordingly. After evaluating the experiments results, Harris Laplace method is a slightly higher recognition rate and the better performance than Difference-of-Gaussian (DoG). The Geometric verification (GV) is also able to increase recognition accuracy rate by filtering out of poor geometric consistency. Future development works may include the expansion of the image database and the improvement of the matching efficiency. |
| first_indexed | 2024-10-01T07:34:23Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/64672 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:34:23Z |
| publishDate | 2015 |
| record_format | dspace |
| spelling | ntu-10356/646722023-07-07T17:32:27Z Location-based services for mobile devices Thin, Mya Mya Yap Kim Hui School of Electrical and Electronic Engineering DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision The explosive growth of mobile technology is widely used for various opportunities in mobile media application development. Mobile apps which involve content-based visual information retrieval allow users to retrieve the information immediately through images. This project aims to transform a mobile device into a portable information terminal that renders image recognition services for mobile devices. This project is to explore and evaluate various image recognition techniques to achieve high performance accuracy with shortest cost (execution time) on stamp image by experimenting the existing image recognition tools. This final year project is for mobile users to be able to identify different stamps of various countries by capturing the images of unknown stamps. In this project, the area of focus is to create a visual stamp database consisting of reference and test databases to perform image recognition process by using the Matlab program. Image recognition tools adopted the Bag-of-Words system and the Scale-Invariant Feature Transform (SIFT) is used for feature extraction. And also, the scalable vocabulary tree and hierarchical k-means are used for machine learning and clustering accordingly. After evaluating the experiments results, Harris Laplace method is a slightly higher recognition rate and the better performance than Difference-of-Gaussian (DoG). The Geometric verification (GV) is also able to increase recognition accuracy rate by filtering out of poor geometric consistency. Future development works may include the expansion of the image database and the improvement of the matching efficiency. Bachelor of Engineering 2015-05-29T04:04:36Z 2015-05-29T04:04:36Z 2015 2015 Final Year Project (FYP) http://hdl.handle.net/10356/64672 en Nanyang Technological University 48 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Thin, Mya Mya Location-based services for mobile devices |
| title | Location-based services for mobile devices |
| title_full | Location-based services for mobile devices |
| title_fullStr | Location-based services for mobile devices |
| title_full_unstemmed | Location-based services for mobile devices |
| title_short | Location-based services for mobile devices |
| title_sort | location based services for mobile devices |
| topic | DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision |
| url | http://hdl.handle.net/10356/64672 |
| work_keys_str_mv | AT thinmyamya locationbasedservicesformobiledevices |