A new segmentation approach for handwritten digits

This article deals with a new segmentation approach applied to unconstrained handwritten digits. The novelty of the proposed algorithm is based on the combination of two types of structural features in order to provide the best segmentation path between connected entities. This method was developed to be applied in a segmentation-based recognition system. In this article, we first present the features used to generate our basic segmentation points. Then, we define our segmentation paths depending on the encountered configurations with only few heuristic rules. Finally, we evaluate the output of our segmenter using a neural network trained with isolated digits.