Detection and Distortion Correction of Arbitrary 2D Barcodes

The development of aesthetically appealing machine-readable codes has progressed significantly, evolving from simple square blocks to sophisticated approaches that integrate shapes or a base image into the code structure. However, detecting and aligning such codes remains a challenge, particularly due to the reliance on algorithms tailored for each code class. This work addresses the need for universality by proposing a detection and rectification framework that operates solely on the structural assumption common to all 2D barcodes: a grid of aligned black and white square blocks.We introduce a multi-resolution detection strategy followed by a robust rectification algorithm. The proposed method complements existing decoding libraries and significantly improves decoding accuracy across various 2D barcode types, including novel and stylized graphic codes. Experimental results demonstrate the framework’s versatility and effectiveness, enabling seamless adaptation to newly designed codes without the need for retraining or reconfiguration.