A New Approach Based on Ncut Clustering Algorithm for Signature Segmentation

In signature recognition, it is a significant step to segment a text line into characters. And the unsupervised clustering is a commonly used remedy for this task. However, due to the strong overlapping and touch among characters, the separation boundaries between two characters are usually nonlinear, which leads to the failure of the widely used clustering methods. To tackle this problem, this paper proposes a new signature segmentation method, which can effectively segment nonlinearly separable characters. In the proposed approach, we first segment the entire signature into strokes, the similarity matrix of which is computed according to stroke gravities. Then, the nonlinear clustering method termed Normalized cut (Ncut) is performed on this similarity matrix to obtain cluster labels for these strokes, through which the strokes are combined into characters. Experiments on four databases are conducted to demonstrate the effectiveness of the proposed method.