Recognition of table-form documents using high order correlation method
Ren-Jean Liou, Mu-Song Chen · 2002
Table-form document recognition has many applications in office automation. An algorithm is proposed in the paper for automatic form processing. A high order correlation method was originally developed for point target detection in three-dimensional space. It computes the spatio-temporal cross-correlations of consecutive data to extract track information in series of images. It was shown that the method provides very good target detection and noise rejection rates. The technique can be easily reformulated in two-dimension for curve detection in regular images. Form processing is one of the good application examples of this technique. In the paper, we apply the high order correlation method to table-form document recognition. We also show that this process is relatively efficient and accurate in detecting and identifying all the lines in a document. In addition, the facts that the algorithm can be implemented using a connectionist network structure further improve the performance. The effectiveness is demonstrated in the simulation results.