Accurate Modelling and Efficient Estimation of the Print-Capture Channel with Application in Barcoding

Area of Science:

• Computer Science
• Electrical Engineering
• Information Theory

Background:

• Mobile cameras are increasingly used for communication, posing challenges for reliable data transmission over print-capture channels.
• Existing print-capture channel models are non-linear and non-stationary, limiting the application of standard channel coding techniques.
• Device-dependent pre-training is often required for current systems, increasing costs and complexity.

Purpose of the Study:

• To develop an accurate parametric model for the print-capture channel.
• To propose an efficient channel estimation scheme with low training overhead.
• To enhance the reliability and data capacity of communication systems utilizing mobile cameras for print-capture tasks.

Main Methods:

• Developed an accurate parametric print-capture channel model.
• Proposed an efficient channel estimation scheme with low training overhead.
• Linearized the print-capture channel to an additive white Gaussian noise (AWGN) model using estimated parameters.
• Applied state-of-the-art channel coding, specifically low-density parity-check (LDPC) codes with iterative soft decoding.

Main Results:

• The proposed model linearizes the print-capture channel to an AWGN channel.
• A demonstrative 8-level 2D barcode using LDPC codes achieved 10% and 56% data capacity gains over QR codes in in-focus and blur-limited scenarios, respectively.
• The channel estimation scheme incurred only a 3.3% additional training overhead.

Conclusions:

• The proposed parametric model and estimation scheme significantly improve communication reliability over print-capture channels.
• Advanced channel coding, like LDPC with soft decoding, can be effectively utilized after channel linearization.
• The developed multilevel 2D barcode demonstrates practical application and enhanced data capacity for reliable mobile camera-based communication.