An Improved Flexible Design with Efficient Detection Algorithm for OILU Code

Artificial tags are widely used in computer vision for object localization and tracking. However, most existing fiducial tags rely on rigid designs that require perfect layouts, limiting their creation and decoding to specialized software. This paper revisits the OILU Code design [1] and proposes a more flexible layout tolerating minor imperfections, allowing manual creation and interpretation by humans and machines. Its pyramidal topology enables the nesting of decimal patterns, making drawing an OILU Code as simple as writing a number. The new layout also supports embedding additional information, such as logos, to meet user-specific needs. An efficient detection method based on cumulative histogram analysis is introduced. Experiments on a large dataset show significant improvements in processing time, robustness to occlusions, and detection range, outperforming other solutions in handling geometric.