State-of-the-Art: Offline Writer Identification Methodologies

This paper reviews various text independent writer identification strategies using offline documents. Various features extraction methodologies are discussed. Handcrafted and Deep Learning based methodologies for classification that are mainly used for offline writer recognition by the researchers and verification by different groups and individuals are presented. Accuracies on identification achieved by the reviewed papers are tabulated and analysed. A survey of different databases used in the reviewed papers is performed. Application of writer identification in different language domains is also discussed. Experimental results of reviewed papers demonstrate that the methodologies get highest accuracy rates till 99.80%.