The language model is a widely used component in fields such as natural language processing, automatic speech recognition, and optical character recognition. In particular, statistical machine translation uses language models, and the translation speed and the amount of memory required are greatly affected by the performance of the language model implementation. We propose a fast and compact implementation of n -gram language models that increases query speed and reduces memory usage by using a double-array structure, which is known to be a fast and compact trie data structure. We propose two types of implementation: one for backward suffix trees and the other for reverse tries. The data structure is optimized for space efficiency by embedding model parameters into otherwise unused spaces in the double-array structure. We show that the reverse trie version of our method is among the smallest state-of-the-art implementations in terms of model size with almost the same speed as the implementation that performs fastest on perplexity calculation tasks. Similarly, we achieve faster decoding while keeping compact model sizes, and we confirm that our method can utilize the efficiency of the double-array structure to achieve a balance between speed and size on translation tasks.

N gram language models tend to increase in size with inflating the corpus size, and consume considerable resources.In this paper, we propose an efficient method for implementing ngram models based on doublearray structures.First, we propose a method for representing backwards suffix trees using double-array structures and demonstrate its efficiency.Next, we propose two optimization methods for improving the efficiency of data representation in the double-array structures.Embedding probabilities into unused spaces in double-array structures reduces the model size.Moreover, tuning the word IDs in the language model makes the model smaller and faster.We also show that our method can be used for building large language models using the division method.Lastly, we show that our method outperforms methods based on recent related works from the viewpoints of model size and query speed when both optimization methods are used.

This paper addresses the problem to recover drawing order from single-stroked offline handwritten image. The recovery problem is formulated as to find the smoothest path to cover all the edges in the graph representation of an input image. The two main contributions are: (1) we introduce direction context to calculate the smoothness between edges; (2) the smoothest path is found by solving a new graph problem: the optimal Euler path problem. An O(n+m+nlog(m/2)) time algorithm is developed to find the optimal Euler path in a graph with 4-degree nodes only. The double traced lines are identified using the maximum weighted matching of general graph. Experimental results on about 13,000 static images converted from the online data in the Unipen database demonstrated the utility of our methods.

Restoration of writing order from a single-stroked handwriting image can be seen as the problem of finding the smoothest path in its graph representation. In this paper, a 3-phase approach to restore a writing order is proposed within the framework of the Edge Continuity Relation (ECR). In the initial, local phase, in order to obtain possible ECRs at an even-degree node, a neural network is used for the node of degree 4 and a theoretical approach is presented for the node of degree higher than 4 by introducing certain reasonable assumptions. In the second phase, we identify double-traced lines by employing maximum weighted matching. This makes it possible to transform the problem of obtaining possible ECRs at odd-degree node to that at even-degree node. In the final, global phase, we find all the candidates of single-stroked paths by depth first search and select the best one by evaluating SLALOM smoothness. Experiments on static images converted from online data in the Unipen database show that our method achieves a restoration rate of 96.0 percent.

Problem of recovering the writing order from single-stroked handwritten image can be seen as finding the smoothest Euler path in its graph representation. In this paper, a novel approach is proposed to solve the recovery problem within the framework of the edge contiguous relation (ECR). Firstly, we make local analyses to obtain the possible ECRs at each of the nodes; secondly a global trace is executed to find all of the candidate Euler paths and the smoothest one is selected as a final result. Based on two simple assumptions, we prove a series of theorems to obtain possible ECRs at even node. Double-traced lines are identified by using the weighted matching of general graph. Experiments on the scanned images and offline images converted from the online data of Unipen database have shown that our method achieved 95.2% correct recovery rate.

Describes a method to recover a drawing order of a handwritten script from a static 2D image. The script should be written in a single stroke and may include double-traced lines. After the script is scanned in and preprocessed, we apply our recovery method which consists of two phases. In the first phase, we globally analyze the graph constructed from the skeletal image and label the graph by determining the types of each edge. In the second phase, we trace the graph from the start vertex to the end vertex using the labeling information. This method does not enumerate the possible cases, for example, by solving the traveling salesman problem and, therefore, does not cause a combinatorial explosion even if the script is very complex. By recovering a drawing order of a handwritten script, the temporal information can be recovered from a static 2D image. Hence, this method will be used as a bridge from the offline handwriting character recognition problem to the online one.

Describes a method to recover the drawing order of a multi-stroke handwritten script from a binary 2D image. First, we construct a graph from the scanned image by applying a thinning process and extracting the skeletal pixels. Next, we identify the start or end vertices in the graph, and then globally analyze the graph to label it by determining the types of each vertex and each edge. Finally, we trace all the strokes using the labeling information and recover the drawing order. The method does not enumerate the possible drawing orders and does not cause a combinatorial explosion, even if the script is very complex. By recovering the drawing order of a handwritten script, the temporal information can be recovered from a scanned image. Hence, this method can be used as a bridge from the offline handwriting character recognition problem to the online one.

A method of handwriting signal encoding based on adaptive linear predictive coding (ALPC) is studied. The ALPC is a form of DPCM which uses a sequentially adaptive predictor in which a sequential estimation algorithm is used to update predictor coefficients. To improve the estimates of the predictor coefficients in the presence of quantization noise, Kalman filtering has been investigated for its feasibility. This results in improvements of not only the estimation of the predictor coefficients, but the signal-to-quantization-noise ratio (SNR) of the signals reconstructed at the receiver as well. Computer simulations have verified that the ALPC system employing the Kalman filter promises high performance and feasibility at the rate of 192 bits/s when applied to handwriting signal encoding.