Abstract. We present a hybrid machine learning approach for information extraction from unstructured documents by integrating a learned classifier based on the Maximum Entropy Mod...
Global and local adaptive thresholding techniques have been shown effective on particular types of documents. None produces consistently good results on all types of documents. In...
International crime and terrorism have drawn increasing attention in recent years. Retrieving relevant information from criminal records and suspect communications is important in...
Decomposition of electromyogram (EMG) provides a valuable means of obtaining motor unit recruitment and firing rate information. The feasibility of decomposing surface EMG signals...