A Comparison between Two Image Detection Algorithms on Neck Angle Detection and a Prolonged Usage Classification Concept
Worawat Lawanontmf15421@shibaura-it.ac.jp
Pornchai Mongkolnampornchai@sit.kmutt.ac.th
Chakarida Nukoolkitchakarida@sit.kmutt.ac.th
Masahiro Inoueinouem@sic.shibaura-it.ac.jp
Graduate School of Engineering and Science, Shibaura Institute of Technology
School of Information Technology, King Mongkut's University of Technology Thonburi, Bangkok, Thailand
Regarding the text neck syndrome, in our previous work we proposed a solution to accurately detect a neck angle while using a smartphone. In this paper, to improve the accuracy and calculation speed, we performed a comparison between two image processing algorithms, which involved in calculating neck angles while using the smartphone. The two image detection algorithms were Haar and LBP (Local Binary Patterns). Both of them had their own advantages and disadvantages. The main difference between the two algorithms was that Haar used floating-point for the calculation, while LBP used integer numbers. The comparison showed the differences of the two algorithms in terms of accuracy and calculation speed. Both Haar and LBP classifiers were trained with 900 positive images and 2,842 negative images. This experiment showed that a combination of both Haar and LBP algorithm had benefits for our system the most. Moreover, for a more effective neck angle detection system, we also proposed a classification of unhealthy neck angle which also concerned with the duration of smartphone usage. This work would encourage the user to have a more healthy neck angle while using the smartphone.
---