Hybrid Optical/Digital Image Processing Applied to Handwriting Recognition and Aerial Photograph Clustering

Writer identification of Chinese handwriting documents using hidden Markov tree model

“…He pointed out that two general approaches had been proposed for the off-line, text-independent writer identification: transform techniques and histogram descriptions. In transform techniques, Duvernoy et al had reported that the most important variations of the writer's transfer function are reflected in the low frequency bands of the Fourier spectrum of handwritten pages [10,11]. Durvernoy et al had also designed a hybrid opticaldigital image processing system to extract features from Fourier spectra of handwritten texts.…”

“…Durvernoy et al had also designed a hybrid opticaldigital image processing system to extract features from Fourier spectra of handwritten texts. The dominant eigenvectors of the data covariance matrix of the Fourier description of the difference between spectras were used as features [11]. Similarly Kuckuck et al had used Fourier transform techniques to preprocess handwritten text as a texture.…”

Holography in museums

“…It can also be used for synthesis viewing of damaged objects or complete presentation of lost objects. Technical holography covers a broader spectrum of applications, including: (1) nondestructive inspection for the detection and visualisation of hidden defects, restoration quality control, object demonstration, storage and preservation; [6][7][8] (2) characterisation and identification of museum objects; 9 (3) high-density data storage and other related usages. This paper discusses some aspects of the practical implementation of these applications.…”