The recovery of handwriting’s dynamic stroke is an effective method to help improve the accuracy of any handwriting’s authentication or verification system. The recovered trajectory can be considered as a dynamic feature of any static handwritten images. Capitalising on this temporal information can significantly increase the accuracy of the verification phase. Extraction of dynamic features from static handwritings remains a challenge due to the lack of temporal information as compared to the online methods. Previously, there are two typical approaches to recover the handwriting’s stroke. The first approach is based on the script’s skeleton. The skeletonisation method has highly computational efficiency whereas it often produces noisy artifacts and mismatches on the resulted skeleton. The second approach deals with the handwriting’s contour, crossing areas and overlaps using parametric representations of lines and thickness of strokes. This method can avoid the artifacts, but it requires complicated mathematical models and may lead to computational explosion. Our paper is based on the script’s extracted skeleton and provides an approach to processing static handwriting’s objects, including edges, vertices and loops, as the important aspects of any handwritten image. Our paper is also to provide analysing and classifying loops types and human’s natural writing behavior to improve the global construction of stroke order. Then, a detailed tracing algorithm on global stroke reconstruction is presented. The experimental results reveal the superiority of our method as compared with the existing ones.

Abstract-In brain-computer interface (BCI) research, there must be a trade-off between accuracy and speed of the BCI system, especially those based on event-related potentials (ERPs). This paper proposes a novel method which can significantly increase the spelling bit rate while also maintaining the desired accuracy. We provide an adaptive real-time stopping method based on the scores of ensemble support vector machine classifiers. We apply a criteria assessment process on the classifiers' scores to dynamically stop the ERP-evoked paradigms at any flashing sequence. Our experiments were conducted on three different P300-Speller data sets (BCI Competition II, BCI Competition III and Akimpech). Our proposed framework significantly outperformed the related state-of-the-art studies in terms of character output accuracy and elicitation bit rate rise between static and dynamic stopping schemes. We improve the average bit rate by over 80% while perfectly maintaining the best original static accuracy of over 96%.

No abstract