Imprinto: Enhancing Infrared Inkjet Watermarking for Human and Machine Perception

TL;DR

Imprinto addresses the challenge of augmenting paper with rich digital content without visual clutter by embedding invisible IR watermarks using IR-absorbing ink printed with off-the-shelf inkjet hardware. A psychophysical study defines conservative invisibility thresholds across background colors, while a CNN-based ML pipeline enables robust, real-time decoding of IR watermarks captured by a universal mobile reader module. The approach leverages the entire document area, including white space, and supports both online (QR-linked) and offline content, yielding high data capacity without compromising aesthetics. Together with an authoring tool and open-source decoding pipeline, Imprinto demonstrates versatile applications from education and security to personal belongings and offline data access, offering a practical, scalable path toward invisible, high-capacity paper augmentation.

Abstract

Hybrid paper interfaces leverage augmented reality to combine the desired tangibility of paper documents with the affordances of interactive digital media. Typically, virtual content can be embedded through direct links (e.g., QR codes); however, this impacts the aesthetics of the paper print and limits the available visual content space. To address this problem, we present Imprinto, an infrared inkjet watermarking technique that allows for invisible content embeddings only by using off-the-shelf IR inks and a camera. Imprinto was established through a psychophysical experiment, studying how much IR ink can be used while remaining invisible to users regardless of background color. We demonstrate that we can detect invisible IR content through our machine learning pipeline, and we developed an authoring tool that optimizes the amount of IR ink on the color regions of an input document for machine and human detectability. Finally, we demonstrate several applications, including augmenting paper documents and objects.

Figures (12)

• Figure 1: Spectral analysis of the inkjet inks.
• Figure 2: User study samples.
• Figure 3: IR ink DTs for each color tested in our experiment, reaching from a mean DT of 81% all the way to 200% IR ink density. The three classes are provided based on participants' ability to detect the IR ink.
• Figure 4: Study procedure. The participant slides the card steadily to the right to reveal the color bar. When participants noticed a color difference between the top (a) and the bottom (b) of the color bar, they were told to report it.
• Figure 5: IR ink computation. (a) We attempt to mathematical model the limits of human perception to conceal IR ink. (b) We use polynomial fitting to estimate the IR capacity based on $lum$ and $K$. (c) The class limits were determined on the histogram of IR ink percentage.