A method (2304) of decoding a QR code having two initially detected finder patterns (2901, 2902; 2911, 2912) is provided. The method forms (2402) a pattern matching template (2700, 2800) based on characteristics of the detected finder patterns and determines (2403) at least one candidate region (2904, 2905; 2913, 2914) about the detected finder patterns. The candidate region is typically based at least on the relative positions of the detected finder patterns. The method detects (2404) a previously undetected third finder pattern of the QR code in the at least one candidate region by correlating content of the candidate region with the pattern matching template. With the identified third finder pattern and each of the two initially detected finder patterns, decoding (2305) the QR code can then be performed. Also disclosed is a method of detecting a two-dimensional code comprising known target features and coded data in an image. The target features comprise a continuous black region, enclosed entirely by a continuous white region, enclosed entirely by a continuous black region. The method performs line-by-line connected component analysis (403) of the image to determine candidate features of interest. The line-by-line connected component analysis aggregates metrics corresponding to candidate features of interest. The method evaluates (404) the aggregated metrics to determine how similar each candidate feature of interest is to the target feature and forms (405) a candidate region of interest from multiple candidate features of interest. Typically the target features include the copyright symbol © and the registered trade mark symbol ®.