Fast location and segmentation of character in annular region based on normalized cross-correlation
Yang Li, Ping Zhang · 2016
Most of the current segmentation algorithms for Character recognition are conducted within a uniform and stable lightening environment. However, the robustness of that algorithm was lost with noise interference and non-uniform and unstable lightening environment in recognition processing. Targeting to these problems in the feature recognition in annular region, a new algorithm that combination of normalized cross-correlation (NCC) and sector scan mode was proposed in this paper. Firstly, The algorithm by using the NCC relative anti noise ability, accurate matching the characteristics of fast matching character regions in a fixed information, and by using sector scan to get the character region, Finally, conversion the annular region by bilinear interpolation located to rectangle region, to achieve fast and stable character segmentation.