Learn from the past – sequentially one-to-one video deblurring network

With the growing availability of hand-held cameras in recent years, more and more images and videos are taken at any time and any place. However, they usually suffer from undesirable blur due to camera shake or object motion in the scene. In recent years, a few modern video deblurring methods are proposed and achieve impressive performance. However, they are still not suitable for practical applications as high computational cost or using future information as input. To address the issues, we propose a sequentially one-to-one video deblurring network (SOON) which can deblur effectively without any future information. It transfers both spatial and temporal information to the next frame by utilizing the recurrent architecture. In addition, we design a novel Spatio-Temporal Attention module to nudge the network to focus on the meaningful and essential features in the past. Extensive experiments demonstrate that the proposed method outperforms the state-of-the-art deblurring methods, both quantitatively and qualitatively, on various challenging real-world deblurring datasets. Moreover, as our method deblurs in an online manner and is potentially real-time, it is more suitable for practical applications.     

Visualization,Discriminability and Applications of Interpretable Saak Features

In this work, we study the power of Saak features as an effort towards interpretable deep learning. Being inspired by the operations of convolutional layers of convolutional neural networks, multi-stage Saak transform was proposed. Based on this foundation, we provide an in-depth examination on Saak features, which are coefficients of the Saak transform, by analyzing their properties through visualization and demonstrating their applications in image classification. Being similar to CNN features, Saak features at later stages have larger receptive fields, yet they are obtained in a one-pass feedforward manner without backpropagation. The whole feature extraction process is transparent and is of extremely low complexity. The discriminant power of Saak features is demonstrated, and their classification performance in three well-known datasets (namely, MNIST, CIFAR-10 and STL-10) is shown by experimental results.     

基于自适应色彩聚类和上下文信息的自然场景文本检测

自然场景文本检测是图像内容分析和理解的重要前提.本文提出一种基于自适应色彩聚类和上下文信息分析的方法,用于检测自然场景图像文本.首先,将层次聚类和参数自学习策略结合,设计一种自适应色彩聚类方法,提取图像中的候选字符.该自适应色彩聚类方法能针对不同图像自动学习权重阈值,有较好的字符召回率.然后,利用文本中字符成行出现的性质,设计一种基于上下文信息的字符验证策略,既能保证较高字符召回率,也能有效移除非文本字符.最后,合并字符构建文本行,并通过后处理得到文本检测结果.在ICDAR2013公共数据集上的实验结果表明：本文分别获得74.17%的召回率,83.40%的准确率和78.52%的F得分.与其他文本检测方法相比,本文获得了较好的文本检测性能,说明本文方法的优越性.     

Space‐Time Warp Curve for Synthesizing Multi‐character Motions

This paper introduces a new motion‐synthesis technique for animating multiple characters. At a high level, we introduce a hub‐sub‐control‐point scheme that automatically generates many different spline curves from a user scribble. Then, each spline curve becomes a trajectory along which a 3D character moves. Based on the given curves, our algorithm synthesizes motions using a cyclic motion. In this process, space‐time warp curves, which are time‐warp curves, are embedded in the 3D environment to control the speed of the motions. Since the space‐time warp curve represents a trajectory over the time domain, it enables us to verify whether the trajectory causes any collisions between characters by simply checking whether two space‐time warp curves intersect. In addition, it is possible to edit space‐time warp curves at run time to change the speed of the characters. We use several experiments to demonstrate that the proposed algorithm can efficiently synthesize a group of character motions. Our method creates collision‐avoiding trajectories ten times faster than those created manually.     

Spotting and Recognition of Consonant-Vowel Units from Continuous Speech Using Accurate Detection of Vowel Onset Points

In this paper, we propose an efficient approach to spotting and recognition of consonant-vowel (CV) units from continuous speech using accurate detection of vowel onset points (VOPs). Existing methods for VOP detection suffer from lack of high accuracy, spurious VOPs, and missed VOPs. The proposed VOP detection is designed to overcome most of the shortcomings of the existing methods and provide accurate detection of VOPs for improving the performance of spotting and recognition of CV units. The proposed method for VOP detection is carried out in two levels. At the first level, VOPs are detected by combining the complementary evidence from excitation source, spectral peaks, and modulation spectrum. At the second level, hypothesized VOPs are verified (genuine or spurious), and their positions are corrected using the uniform epoch intervals present in the vowel regions. The spotted CV units are recognized using a two-stage CV recognizer. Two-stage CV recognition system consists of hidden Markov models (HMMs) at the first stage for recognizing the vowel category of a CV unit and support vector machines (SVMs) for recognizing the consonant category of a CV unit at the second stage. Performance of spotting and recognition of CV units from continuous speech is evaluated using Telugu broadcast news speech corpus.     

嵌入深度信念网络的点过程模型用于关键词检出

基于点过程模型的关键词检出系统是一种新颖的连续语音关键词检出系统,虽然该系统具有对样本数要求不高、计算速度快等优点,但其检出性能比较依赖于前端音素探测器的准确度,而目前广泛用于音素探测器的高斯混合模型存在表征和建模能力不强的问题。针对这一缺陷,本文提出了一种嵌入深度信念网络的点过程模型并将其应用于关键词检出,该模型采用表征能力强的深度信念网络来建立音素探测器,改进了高斯混合模型在表征能力上的不足。实验结果表明该方法能够获得比原模型更高的检出率,并且降低了计算复杂度,更适用于需要实时检测关键词的场合。      

基于加权有限状态机的动态匹配词图生成算法

由于现有的加权有限状态机(WFST)解码网络没有精确词尾标记,导致当前已有的词图生成算法不含精确的词尾时间点,或者仅是状态、音素级别的词图,无法应用到关键词检索中。该文提出在WFST静态解码器下的语音识别词图生成算法。首先从理论上分析了WFST解码音素图和词图的可转换关系,然后提出了字典的动态音素匹配方法解决了WFST网络中词尾时间点对齐的问题,最后通过令牌传递的遍历方法生成了词图。同时,考虑到计算量优化,在令牌传递过程中引入了剪枝算法,使音素图转词图的耗时不到解码耗时的3%。得到的词图,不仅可以用于语言模型重打分,由于含有精确的词尾时间点,还可以直接应用到关键词检索系统中。实验结果表明,该文的词图生成算法具有较高的计算效率;和已有动态解码器的词图相比,词图中包含更多解码信息,在大词汇连续语音识别的重打分结果和关键词检索中都能取得更好的性能。     

Hybrid digital-analog technique for UWB direction finding

In this letter, we propose an alternative approach for UWB direction finding using hybrid digital-analog technique. The proposed system uses an array of leading-edge pulse detection scheme in analog domain. The detection scheme is realized using tunnel diode with envelope detector and comparator latch, whose output is digital binary signal. The outputs or all array branches are then summed up to form a staircase-shaped waveform, whereby the slope of the waveform indicates the direction of impinging UWB signal. In order to detect the slope and produce the estimation of the direction, an analog differentiator joint with a peak detector is utilized. The proposed system is relatively simple and does not require high sampling rate or high speed DAC.     

A NOVEL DETECTOR FOR UPLINK CDMA SYSTEM WITH UNKNOWN INTERFERENCE

In this letter, the detection of asynchronous DS-CDMA signal with multipath fading and interference from neighboring cells is studied. A novel multiuser detector based on Gibbs sampler is proposed, in which Gibbs sampler is employed to perform the Bayesian multiuser detection according to the linear group-blind decorrelator output. Since Gibbs sampler is dependent of parameter estimation that can be improved by the output of the detector, an enhanced Gibbs sampier based detector using the improved parameter estimation is put forward. The novel multiuser detection technique has the advantages of high performance and wide applications. Computer simulations show its effectiveness.     

A novel maneuver detector based on back propagation neural network

Target maneuvering is always accompanied with rapid attitude variations, which are helpful to achieve high cross-range resolution for coherent pulse radar. In this paper, the high resolution Doppler profile (HRDP) is first formulated. The principle of maneuver detection using HRDP is then fully exploited. The difference of target attitude rates between nonmaneuvering and maneuvering motion mode is analyzed. Due to the nonstationarity of HRDP, the maneuver detection problem is reformulated as a pattern classification problem, where nonmaneuvering and maneuvering motion mode are distinguished. A novel detector is then developed based on the back propagation neural network. Two novel indices for performance evaluation are proposed. They reflect the dynamic performance of the maneuver detector more reasonably than the classical index, average detection delay. Finally, the simulation results show that the proposed detector possesses low detection delay and high detection probability.     

Improved Robust Techniques for Multiuser Detection in Non-Gaussian Channels

In many physical channels where multiuser detection techniques are to be applied, the ambient channel noise is known through experimental measurements to be decidedly non-Gaussian, due largely to impulsive phenomena. This is due to the impulsive nature of man-made electromagnetic interference and a great deal of natural noise. This paper presents a robust multiuser detector for combating multiple access interference and impulsive noise in code division multiple access (CDMA) communication systems. A new M-estimator is proposed for "robustifying" the detector. The approach is corroborated with simulation results to evaluate the performance of the proposed robust multiuser detector compared with that of the linear decorrelating detector, and the Huber and the Hampel M-estimator based detectors. Simulation results show that the proposed detector with significant performance gain outperforms the linear decorrelating detector, and the Huber and the Hampel M-estimator based detectors. This paper also presents an improved robust blind multiuser detection technique based on a subspace approach, which requires only the signature waveform and the timing of the desired user to demodulate that user's signal. Finally, we show that the robust multiuser detection technique and its blind adaptive version can be applied to both synchronous and asynchronous CDMA channels.     

高速超宽带脉冲检测电路的设计

在介绍了一种基于TH-PPM调制方式的并行检测接收机方案的基础上,结合UWB脉冲信号的时域特征提出并实现了一种结构简单、性能稳定的脉冲能量检测电路.实验结果表明该电路可在100Mb/s的高传输速率情况下取得较好的超宽带信号检测效果.     

Hierarchical deep belief networks based point process model for keywords spotting in continuous speech

Point process model keyword spotting (KWS) system has attracted considerable attentions in the areas of keyword spotting by its capacity that can generalize from a relatively small numbers of training examples. But unfortunately, the accuracy level of the point process model is not comparable with the state‐of‐the‐art KWS systems because of the poor modeling capacity of the phoneme detector, which are based on Gaussian Mixture Models. In this paper, focus on improving the performance of detector in point process model, we propose an enhanced version of point process model, which is based on hierarchical deep belief networks (DBNs). Hierarchical DBNs are used as the phoneme detector in this system, and they combine the advantages of both the DBN and the hierarchical architecture for capturing complex statistical patterns in speech while overcoming the inherent flaws of conventional hidden Markov models and multilayer layer perceptron. Experiments results on TIMIT database show that the proposed method can yield 2% improvement. Furthermore, in the case when training examples are extremely limited, it can achieve better results over state‐of‐the‐art KWS systems. Copyright © 2013 John Wiley & Sons, Ltd.     

基于Gibbs采样的群盲多用户检测技术

多用户检测是第三代移动通信系统中的一项关键技术。研究存在邻近小区干扰时异步多径DS-CDMA信号的多用户检测问题。提出一种新颖的基于Gibbs采样的群盲多用户检测技术，它根据线性群盲解相关器的输出，利用Gibbs采样进行Bayes多用户检测。存在信道编码时，提出的检测器非常适合做Turbo多用户检测，仍然能够找到Bayes解。该方法具有简单易实现、检测性能好、应用范围广泛的优点。仿真实验证实了这一点。     

Principal visual word discovery for automatic license plate detection

License plates detection is widely considered a solved problem, with many systems already in operation. However, the existing algorithms or systems work well only under some controlled conditions. There are still many challenges for license plate detection in an open environment, such as various observation angles, background clutter, scale changes, multiple plates, uneven illumination, and so on. In this paper, we propose a novel scheme to automatically locate license plates by principal visual word (PVW), discovery and local feature matching. Observing that characters in different license plates are duplicates of each other, we bring in the idea of using the bag-of-words (BoW) model popularly applied in partial-duplicate image search. Unlike the classic BoW model, for each plate character, we automatically discover the PVW characterized with geometric context. Given a new image, the license plates are extracted by matching local features with PVW. Besides license plate detection, our approach can also be extended to the detection of logos and trademarks. Due to the invariance virtue of scale-invariant feature transform feature, our method can adaptively deal with various changes in the license plates, such as rotation, scaling, illumination, etc. Promising results of the proposed approach are demonstrated with an experimental study in license plate detection.     

License Plate Recognition System Using Artificial Neural Networks

A high performance license plate recognition system (LPRS) is proposed in this work. The proposed LPRS is composed of the following three main stages: (i) plate region determination, (ii) character segmentation, and (iii) character recognition. During the plate region determination stage, the image is enhanced by image processing algorithms to increase system performance. The rectangular license plate region is obtained using edge‐based image processing methods on the binarized image. With the help of skew correction, the plate region is prepared for the character segmentation stage. Characters are separated from each other using vertical projections on the plate region. Segmented characters are prepared for the character recognition stage by a thinning process. At the character recognition stage, a three‐layer feedforward artificial neural network using a backpropagation learning algorithm is constructed and the characters are determined.     

基于并行遗传算法的CDMA优化多用户检测方法

多用户检测技术是第三代移动通信系统CDMA中的一项关键技术。在多用户检测中求解最佳矢量问题可以转化为在遗传算法中求解具有最高适应度函数的问题。本文提出了一种基于并行遗传算法的CDMA多用户检测器,并与最佳多用户检测和传统检测器进行比较,实验结果表明本方法可获得接近最佳检测的性能。由于采用并行遗传算法,这种多用户检测器更易于实时应用和硬件实现。     

自然场景下的车牌检测与识别算法

车牌自动识别是智能交通系统的关键技术之一,主要包括车牌检测和字符识别两部分。为提高车牌检测速度和精度,本文提出了一种基于学习、由粗到精的车牌检测方法。首先采用颜色点对和垂直边缘相结合的方法,快速检测出车牌感兴趣区域;然后采用一种基于梯度方向直方图特征和支持向量机的机器学习方法实现车牌的精确定位。在车牌识别阶段,首先采用基于连通域分析与字符固有特征相结合的方法进行字符分割,然后根据字符结构提取3种稳定且有效的特征,采用支持向量机对分割的字符进行识别。采用上述方法对412幅不同角度、不同光照条件、不同时间段下拍摄的图像进行检测与识别,实验结果表明本文提出的算法精度高、鲁棒性好、识别速度符合实时性的要求。