Hosted by Utah State University Libraries
On Mobile Detection and Localization of Skewed Nutrition Facts Tables
Abstract
dc:description.abstractThe Computer Science Assistive Technology Laboratory (CSATL) at Utah State University has a long history of research in visually impaired grocery shopping tech. CSATL’s ShopMobile II introduced nutrition facts table (NFT) analysis but only with perfectly aligned and square input images. A new method which detects and localizes NFTs more quickly and from rotated or non-square images has been released and is slated for integration with ShopMobile II to improve this feature substantially. This is great news for the estimated 3.6 million adults in the United States having visual impairment or blindness and also opens the doors to other applications where analysing NFTs can significantly aid users. By combining image analysis methods in creative ways, the new method avoids detecting NFTs where there are none and properly locates them in about 42% of images. This is remarkable considering images will be processed as quickly as possible on the device – a standard Android smartphone. The CSATL isn’t stopping here though: the new method exposes several possibilities for further improvements.
Degree
thesis:*- Name thesis:degree_name
- Master of Science (MS)
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Computer Science
- Grantor dc:publisher
- Hosted by Utah State University Libraries
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Blay, Christopher
- Contributors dc:contributor
-
- Vladimir A. Kulyukin
- Curtis Dyreson
- Nick Flann