Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-01-03 Cooperative journals: 《计算机应用研究》
Abstract： In the off-line signature, LBPs of background pixels, as well as pixels inside the strokes, accounting for a large proportion in a signature bitmap, are nearly the same. Consequently, they have greate interference to describe the characteristics of signature’s strokes. So this paper proposed LBPC-based feature that mainly calculated the histogram of the local binary patterns on signature’s contour and some useless patterns were removed by adding a rule to improve the effectiveness and robustness of LBP. Besides, seen as an improvement for directional chain code for its limitations in the signature verification, this paper introduced LCPC-based feature which aimed at computing the statistical features of the local contour patterns co-occurrence. Then, the PCA was applied to two above combined features due to the huge dimensionality. Finally, to evaluate the performance of proposed method, using SVMs classifier, it conducted experiments on MCYT and GPDS open databases and the achievable average error rates were 13.51% and 12.97% respectively. Moreover, comparisons with other methods on the same datasets provide evidence that the proposed method obtains lower average error rate than others.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-24 Cooperative journals: 《计算机应用研究》
Abstract： In order to improve unregistered word recognition effect in various fields, this paper proposed an unregistered word recognition method based on expansion rules and statistical features. It analyzed word formation features of unregistered words in various field, formulated expansion rules, extended word segmentations to get compound words according to expansion rules, then determined whether compound words were unregistered words through statistical features such as word frequency, mutual information and branch entropy, if the compound word was an unregistered word, it would continue to be expanded and recognized. The results of unregistered word recognition experiments in six fields and general field show that the method based on expansion rules and statistical features achieves better recognition effect of unregistered words and has better portability.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-20 Cooperative journals: 《计算机应用研究》
Abstract： The sequence of the in-air signature is long, in order to make the signature match faster and take advantage of more local information, this paper proposed a method that use the extreme points to segment the signature data and then measure the distance. Aiming at the mismatch problem of different polarity extreme point in traditional DTW algorithm, it proposed an improved DTW algorithm based on extreme point matching, which constrained the matching path of DTW algorithm and avoided error matching between extreme points. It achieved the result FRR: 4.15% and FAR: 3.82% on the local database. Experimental results show that, compare with the traditional global distance measurement method, the method that segment first and then measure the distance improves the accuracy and efficiency of the system.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-02 Cooperative journals: 《计算机应用研究》
Abstract： In order to reduce the missing detection and misclassification of text caused by uneven illumination and background complexity in text detection of natural scenes, this paper presented a natural scene text detection method based on stroke angle transformation and width features. Compared to non-text, the text has a more stable performance of stroke outline angle conversion times and stroke width. Therefore, this paper proposed methods of extracting the number of transformations of the outer corner of the stroke and the enhancement of the pixel area ratio of the stroke support. In order to extract the characteristics of angular conversion, it used the method of outer contour segmentation to calculate the number of conversion times. In order to extract the strokes width characteristics, it calculated the proportion of the width stable area in the total strokes area. To reduce rate of the text missing detection, multi-channel MSER was used to detect text candidate area. Candidate areas in all channels were merged to extract the stroke and texture features. Support vector machines combined with features adopted, it used to classify text and non-text. The simulations show that the accuracy and recall rate of the algorithm were 79.3% and 72.8% in the ICDAR2015 database, respectively. Moreover, it solves the problem of uneven illumination and complex background to some extent.
Hits
Hits
Downloads
Comment