Mobile visual object search and recognition

• Main Author: Wong, Li Yan
• Other Authors: Yuan Junsong
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/67689
• Institution: Nanyang Technological University

The proliferation of indoor positioning and navigation mobile applications based on various technologies, such as Wi-Fi access points, Bluetooth, Infrared, Radio Frequency Identification (RFID) and Near Field Communication (NFC) for indoor estimation positioning have already been implemented and are available on the market. In this project, we explore the feasibility of determining pedestrian’s location based on the images captured by a smart phone. Image Process/Retrieval API for Object Recognition would do the matching of captured image and retrieval of images in an offline database that is computed and stored in smart phone. The matching is done by features comparison between the captured image and database’s images. When there is a match between captured image and retrieval of images in database found, the captured image would do a lookup in an XML file to find a match between the name of database image and name of image stored in the XML file. Information such as name of place, floor, floor map and location would be returned with location marker identifying the pedestrian’s location on the 2-Dimensional floor map.