This work presents a first contribution to the discrimination of the medieval manuscript texts that helps palaeographers to date the ancient manuscripts.

This work should confirm objectively to the work of the palaeographers. It tests also the possibility of discriminating the medieval writings by image analysis.



The principle idea of our work is to characterise the shape of writings, without any segmentation. Our method is based on the Spatial Grey-Level Dependence (SGLD) which measures the joint probability between grey levels values of pixels for each displacement.



We have shown that the SGLD generalises many other auto similarity measures. The SGLD satisfies many conditions that we have fixed from the beginning. We use the Haralick features to reduce the number of descriptors which characterise the SGLD matrices for each image.

The achieved discrimination results are encouraging and confirming to the classifications of writing given by palaeographers. More interesting results are obtained by a new spatial dependence based on orientations, curvature that we have defined.