AR Image Target Generator
A tool designed to simplify the creation of high-quality image markers for augmented reality applications.
Finding suitable images with good tracking performance for marker-based AR/MR projects can be time-consuming and challenging. Images need to have high contrast, a high feature count, and a homogeneous feature distribution to ensure reliable tracking.
In 2023, during my master's thesis on Augmented Reality Task Assistance in Life Science Laboratories, while developing an AR application for the HoloLens 2 using the Vuforia SDK's Image Target feature, I encountered this very problem.
Fortunately, I came across a blog post by Dario Mazzanti from 2014, which provided a Processing sketch to generate markers meeting all these requirements. With some tweaking, I was able to create a marker that matched my desired aspect ratio and incorporated my product colors.
While this code was helpful, I had to spend time understanding the marker generation algorithm and how to tweak it. Additionally, I needed to download and install Processing and re-run the program every time I wanted to generate and save a single marker. This process was cumbersome and time-consuming.
This experience inspired me to develop this AR Image Target Generator, making marker generation and customization easily accessible. Whether you're working on quick prototypes or larger projects requiring multiple markers across various scenes, this tool helps you create high-quality image targets effortlessly. You can customize markers with random colors or align them with your brand identity using your specific product colors—all without needing to get into code or graphic design programs.
A quick how-to guide:
- Set the pixel size of the longest image edge. (Height or Width)‚
- Choose the aspect ratio of your markers. (1:1, 4:3, 16:9)
- Choose the border-style. (None, Black, White)
- Choose between three graphic styles (Mazzanti, Chaos, Dot Matrix)
- Choose a background color. (Black or white, only relevant for Chaos and Dot Matrix style)
- Randomize the colors or use your own colors for the marker.
In the results section, you can see the number of features detected in the generated marker using the OpenCV ORB algorithm. This metric provides an estimate of the marker's quality and tracking performance, allowing you to effectively compare different generated markers.
Acknowledgements:
I would like to express my gratitude to Dario Mazzanti for sharing his insightful blog post and Processing sketch with the public. His work enabled me to conduct my AR research and significantly influenced the creation of this tool. In honor of his contribution, the "Mazzanti" style option in this generator is named after him, as its implementation incorporates elements of his original Processing sketch.