Why not use smartphones/tablets - Semantic Scholar
TOWARDS NEXT GENERATION BARCODE SCANNING Sörös G., Flörkemeier C. - ETH Zurich {soeroesg, floerkem}@inf.ethz.ch
Abstract Smartphones and tablets are increasingly used to scan visual codes that act as physical hyperlinks to digital information [1]. Compared with the outstanding performance of enterprise laser scanners, smartphone cameras suffer from motion/shake blur and limited image resolution. In this project, we propose to turn every smartphone into an enterprise-grade barcode scanner by adapting the latest research results in space-time super resolution to the barcode domain.
Motivation
Why use smartphones/tablets as barcode scanners? everyone carries a smartphone, tablets are getting popular most smartphones have (autofocus) cameras outstanding (parallel) computing and sensing capabilities easy application development and business process integration new enterprise use cases (e.g. tablet as point of sale for small businesses)
Why not use smartphones/tablets as barcode scanners? suffer from various types of blur (e.g., out-of-focus blur, motion/shake blur) no localization cue from laser limited image resolution
Enterprise barcode scanners are expensive devices and therefore only selected people are using them (e.g., airport baggage controllers). Our goal is to improve business processes by providing a ubiquitous barcode scanning solution for each employee in the value chain.
References [1] R. Adelmann - Mobile phone based interaction with everyday products - On the go, in IEEE Int. Conf. on Next Generation Mobile Applications, Services and Technologies, Cardiff, UK, 2007 [2] E. Shechtman, Y. Caspi, M. Irani - Spacetime super-resolution, in IEEE Trans. on Pattern Analysis and Machine Intelligence, vol.27, no.4, pp.531–545, 2005 [3] H. Takeda, P. Milanfar - Removing motion blur with space-time processing, in IEEE Trans. on Image Processing, vol.20, no.10, p.2990, 2011
Acknowledgement Our idea was inspired by [2] and [3]. Photos were taken in FoodLab.
Towards space-time super resolution of barcodes Image deblurring has been an active research area in the past decade, but all previous work made assumptions (e.g., special hardware, uniform point spread function (PSF), static and/or planar scene, or motion segmentation) that are violated in barcode scanning. Recently, Takeda et al. introduced a space-time super resolution [2] method called 3D steering kernel regression [3]. Besides spatial upsampling, the method allows for the removal of motion blur using a shift-invariant spacetime (3D) PSF without any motion information. The algorithm provides outstanding results but is not suitable for real-time processing. Our contribution will be to adapt the general method to the very special properties of barcode images with the goal of super resolution and deblurring of barcode images in real time. The special properties of barcode scanning include: barcodes do not need to be perfectly sharp for decoding intra-frame and inter-frame motions are small the original scene is often planar and contains sharp edges smartphones are equipped with inertial sensors problem is suitable for parallel processing (GPU)
Abstract Smartphones and tablets are increasingly used to scan visual codes that act as physical hyperlinks to digital information [1]. Compared with the outstanding performance of enterprise laser scanners, smartphone cameras suffer from motion/shake blur and limited image resolution. In this project, we propose to turn every smartphone into an enterprise-grade barcode scanner by adapting the latest research results in space-time super resolution to the barcode domain.
Motivation
Why use smartphones/tablets as barcode scanners? everyone carries a smartphone, tablets are getting popular most smartphones have (autofocus) cameras outstanding (parallel) computing and sensing capabilities easy application development and business process integration new enterprise use cases (e.g. tablet as point of sale for small businesses)
Why not use smartphones/tablets as barcode scanners? suffer from various types of blur (e.g., out-of-focus blur, motion/shake blur) no localization cue from laser limited image resolution
Enterprise barcode scanners are expensive devices and therefore only selected people are using them (e.g., airport baggage controllers). Our goal is to improve business processes by providing a ubiquitous barcode scanning solution for each employee in the value chain.
References [1] R. Adelmann - Mobile phone based interaction with everyday products - On the go, in IEEE Int. Conf. on Next Generation Mobile Applications, Services and Technologies, Cardiff, UK, 2007 [2] E. Shechtman, Y. Caspi, M. Irani - Spacetime super-resolution, in IEEE Trans. on Pattern Analysis and Machine Intelligence, vol.27, no.4, pp.531–545, 2005 [3] H. Takeda, P. Milanfar - Removing motion blur with space-time processing, in IEEE Trans. on Image Processing, vol.20, no.10, p.2990, 2011
Acknowledgement Our idea was inspired by [2] and [3]. Photos were taken in FoodLab.
Towards space-time super resolution of barcodes Image deblurring has been an active research area in the past decade, but all previous work made assumptions (e.g., special hardware, uniform point spread function (PSF), static and/or planar scene, or motion segmentation) that are violated in barcode scanning. Recently, Takeda et al. introduced a space-time super resolution [2] method called 3D steering kernel regression [3]. Besides spatial upsampling, the method allows for the removal of motion blur using a shift-invariant spacetime (3D) PSF without any motion information. The algorithm provides outstanding results but is not suitable for real-time processing. Our contribution will be to adapt the general method to the very special properties of barcode images with the goal of super resolution and deblurring of barcode images in real time. The special properties of barcode scanning include: barcodes do not need to be perfectly sharp for decoding intra-frame and inter-frame motions are small the original scene is often planar and contains sharp edges smartphones are equipped with inertial sensors problem is suitable for parallel processing (GPU)
