An automated handwritten signature detection approach for e-security purposes

In this work, a method for off-line handwritten signature compression and verification is proposed. We address two important problems regarding signature — security and memory occupancy (by the signature image). The paper has two contributions — the first contribution is a new frame work for compressing the signature image into an unique one dimensional array, thereby reducing the memory occupancy and the second contribution is comparison of the arrays(signatures) using a probabilistic divergence measure. The proposed mechanism is memory economic, computationally advantageous and shows higher accuracy than the existing ones.