Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216959
Yoshihiro Sugaya, K. Motoyama, S. Omachi
Peer-to-peer system is a promising solution to manage a large amount of data, but similarity search on peer-to-peer network with a restricted small number of messages is a challenging problem. Existing methods that can perform similarity search work only with low-dimensional data. We propose a method to transform the very high-dimensional data into low-dimensional vectors in order to perform similarity search.
{"title":"Data arrangement and dimensional compression using Vivaldi for similarity search on structured peer-to-peer network","authors":"Yoshihiro Sugaya, K. Motoyama, S. Omachi","doi":"10.1109/ICCE-TW.2015.7216959","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216959","url":null,"abstract":"Peer-to-peer system is a promising solution to manage a large amount of data, but similarity search on peer-to-peer network with a restricted small number of messages is a challenging problem. Existing methods that can perform similarity search work only with low-dimensional data. We propose a method to transform the very high-dimensional data into low-dimensional vectors in order to perform similarity search.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"6 35","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120927891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216923
Yen-Teh Liu, Yu Tsao, Ronald Y. Chang
In this paper, we study the problem of Mandarin consonant/vowel separation which is an integral part of many Mandarin speech applications. We propose a deep neural network (DNN) based approach and compare its performance with the support vector machine (SVM) method. Our results demonstrate an improved separation performance yielded by the proposed method, especially on consonant identification.
{"title":"A deep neural network based approach to mandarin consonant/vowel separation","authors":"Yen-Teh Liu, Yu Tsao, Ronald Y. Chang","doi":"10.1109/ICCE-TW.2015.7216923","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216923","url":null,"abstract":"In this paper, we study the problem of Mandarin consonant/vowel separation which is an integral part of many Mandarin speech applications. We propose a deep neural network (DNN) based approach and compare its performance with the support vector machine (SVM) method. Our results demonstrate an improved separation performance yielded by the proposed method, especially on consonant identification.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127057726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216999
C. Yeh, Pin-Hsian Liu, Cheng-En Yu, Chih-Yang Lin
There are many outdoor vision applications such as surveillance and navigation. One of the challenges is rain removal, especially the rain removal from a single image. In this paper, a single rain image is divided into the high frequency part and the low frequency part by the Gaussian filter. Non-negative matrix factorization (NMF) is used to remove the rain streaks in the low frequency part. Then, Canny edge detection is applied to deal with the rain in the high frequency and the block copy method is employed to preserve the image quality. After that, we applied a rain dictionary to further divide the high frequency into rain and non-rain parts. The experimental results show that the proposed method is better than the state-of-the-art methods, especially in the high frequency part.
{"title":"Single image rain removal based on part-based model","authors":"C. Yeh, Pin-Hsian Liu, Cheng-En Yu, Chih-Yang Lin","doi":"10.1109/ICCE-TW.2015.7216999","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216999","url":null,"abstract":"There are many outdoor vision applications such as surveillance and navigation. One of the challenges is rain removal, especially the rain removal from a single image. In this paper, a single rain image is divided into the high frequency part and the low frequency part by the Gaussian filter. Non-negative matrix factorization (NMF) is used to remove the rain streaks in the low frequency part. Then, Canny edge detection is applied to deal with the rain in the high frequency and the block copy method is employed to preserve the image quality. After that, we applied a rain dictionary to further divide the high frequency into rain and non-rain parts. The experimental results show that the proposed method is better than the state-of-the-art methods, especially in the high frequency part.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125493001","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7217039
Do-Hyung Kim, Seok-hwan Jo, K. Kwon, Yeonbok Lee, Seung-Won Lee, Young-Hwan Park, Sukjin Kim, Jaehyun Kim, Shihwa Lee
We introduce an ultra-low-power digital signal processor (DSP) solution for wearable applications with high performance. It employs three-issue VLIW architecture with the major low-power techniques and implemented with 95K gates in Samsung 28LPP process and runs up to 200MHz. The experimental results demonstrate that a voice trigger application can operate at 6.1MHz under 0.15mW power consumption.
{"title":"Ultra-low-power voice trigger for wearable devices","authors":"Do-Hyung Kim, Seok-hwan Jo, K. Kwon, Yeonbok Lee, Seung-Won Lee, Young-Hwan Park, Sukjin Kim, Jaehyun Kim, Shihwa Lee","doi":"10.1109/ICCE-TW.2015.7217039","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7217039","url":null,"abstract":"We introduce an ultra-low-power digital signal processor (DSP) solution for wearable applications with high performance. It employs three-issue VLIW architecture with the major low-power techniques and implemented with 95K gates in Samsung 28LPP process and runs up to 200MHz. The experimental results demonstrate that a voice trigger application can operate at 6.1MHz under 0.15mW power consumption.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127544497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216929
Shih-Shinh Huang, Chun-Yuan Chen
Pedestrian detection and counting is an important topic in developing an intelligent surveillance system. In this work, we propose a vision-based system for detecting pedestrians in an image. Be robust to crowded scenes and adapt to incomplete foreground from background subtraction algorithm, expectation maximization (EM) algorithm is applied to impose the constraint of body part for achieving successful detection. A well-known dataset called CAVIAR is used to validate the effectiveness of the proposed method.
{"title":"Vision-based crowded pedestrian detection","authors":"Shih-Shinh Huang, Chun-Yuan Chen","doi":"10.1109/ICCE-TW.2015.7216929","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216929","url":null,"abstract":"Pedestrian detection and counting is an important topic in developing an intelligent surveillance system. In this work, we propose a vision-based system for detecting pedestrians in an image. Be robust to crowded scenes and adapt to incomplete foreground from background subtraction algorithm, expectation maximization (EM) algorithm is applied to impose the constraint of body part for achieving successful detection. A well-known dataset called CAVIAR is used to validate the effectiveness of the proposed method.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127608143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216989
Po-Chang Wu, Bin-Da Liu, C. Yeh, S. Tseng, H. Tsai, Y. Juang
This paper presents a low-voltage low-power monolithic complementary metal-oxide-semiconductor (CMOS) micro-electromechanical-system (MEMS) accelerometer design. This design utilizes low-voltage design techniques without using low-threshold devices or internal supply voltage boosting. The accelerometer, designed in the 0.18-μm CMOS MEMS process, contains the micro-mechanical structure, readout circuits, and a 16-bit delta-sigma analog-to-digital converter (ΔΣ ADC). It occupies an area of only 0.8 × 1 mm2 and draws 0.33 mA of current from a 0.6-V supply. The simulated sensitivity is 3000 LSB/g and the nonlinearity is 0.78% within the ±6 g sensing range.
{"title":"Design of a 0.6-V 0.2-mW CMOS MEMS accelerometer","authors":"Po-Chang Wu, Bin-Da Liu, C. Yeh, S. Tseng, H. Tsai, Y. Juang","doi":"10.1109/ICCE-TW.2015.7216989","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216989","url":null,"abstract":"This paper presents a low-voltage low-power monolithic complementary metal-oxide-semiconductor (CMOS) micro-electromechanical-system (MEMS) accelerometer design. This design utilizes low-voltage design techniques without using low-threshold devices or internal supply voltage boosting. The accelerometer, designed in the 0.18-μm CMOS MEMS process, contains the micro-mechanical structure, readout circuits, and a 16-bit delta-sigma analog-to-digital converter (ΔΣ ADC). It occupies an area of only 0.8 × 1 mm2 and draws 0.33 mA of current from a 0.6-V supply. The simulated sensitivity is 3000 LSB/g and the nonlinearity is 0.78% within the ±6 g sensing range.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"421 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132555052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216977
M. Muzammil, I. Ali, M. Sharif, K. A. Khalil
Motion Estimation (ME) is the most critical and complex part of any video codec system. The different algorithms and their architectures are proposed for ME process. In this paper, we have proposed an efficient architecture for Hexagon Based Search (HexBS) algorithm and implemented on XC4VSX25 Virtex4 FPGA. Simulation results show that the proposed architecture is capable of calculating the Motion Vectors (MVs) of 1280×720 High Definition (HD) videos with the best case throughput of 70 frames/sec. Moreover, the power and frequency requirements are 215mW and 127.27 MHz respectively for the proposed architecture with minimum hardware resources. Hence the proposed architecture is suitable for the real-time HD video applications.
运动估计是视频编解码系统中最关键、最复杂的部分。针对ME过程提出了不同的算法及其体系结构。本文提出了一种高效的Hexagon Based Search (HexBS)算法架构，并在XC4VSX25 Virtex4 FPGA上实现。仿真结果表明，该架构能够计算1280×720高清视频的运动矢量(mv)，最佳吞吐量为70帧/秒。在硬件资源最少的情况下，该架构的功耗和频率要求分别为215mW和127.27 MHz。因此，所提出的体系结构适用于实时高清视频应用。
{"title":"An efficient FPGA architecture for hardware realization of hexagonal based motion estimation algorithm","authors":"M. Muzammil, I. Ali, M. Sharif, K. A. Khalil","doi":"10.1109/ICCE-TW.2015.7216977","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216977","url":null,"abstract":"Motion Estimation (ME) is the most critical and complex part of any video codec system. The different algorithms and their architectures are proposed for ME process. In this paper, we have proposed an efficient architecture for Hexagon Based Search (HexBS) algorithm and implemented on XC4VSX25 Virtex4 FPGA. Simulation results show that the proposed architecture is capable of calculating the Motion Vectors (MVs) of 1280×720 High Definition (HD) videos with the best case throughput of 70 frames/sec. Moreover, the power and frequency requirements are 215mW and 127.27 MHz respectively for the proposed architecture with minimum hardware resources. Hence the proposed architecture is suitable for the real-time HD video applications.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132723969","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216815
Z. Moghaddasi, H. Jalab, R. M. Noor
Digital image forgery is becoming easier to perform because of the rapid developments of various manipulation tools. Between the various image forgery techniques, image splicing is considered as one the most prevalent technique. In this paper, a low dimensional singular value decomposition (SVD) based feature extraction method applied in steganalysis is proposed as an image splicing detection method. The SVD-based features are applied in different spatial and frequency domains to make a comprehensive comparison between these various transforms. Support vector machine is used to distinguish between authentic and spliced images. The results are encouraging and show that the detection accuracy of 77.60% is achieved for the DCT transform with only 25 dimensional feature vector.
{"title":"A comparison study on SVD-based features in different transforms for image splicing detection","authors":"Z. Moghaddasi, H. Jalab, R. M. Noor","doi":"10.1109/ICCE-TW.2015.7216815","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216815","url":null,"abstract":"Digital image forgery is becoming easier to perform because of the rapid developments of various manipulation tools. Between the various image forgery techniques, image splicing is considered as one the most prevalent technique. In this paper, a low dimensional singular value decomposition (SVD) based feature extraction method applied in steganalysis is proposed as an image splicing detection method. The SVD-based features are applied in different spatial and frequency domains to make a comprehensive comparison between these various transforms. Support vector machine is used to distinguish between authentic and spliced images. The results are encouraging and show that the detection accuracy of 77.60% is achieved for the DCT transform with only 25 dimensional feature vector.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"218 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134311806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216925
Hsin-Ju Hsieh, J. Hung
In this paper, a novel method is proposed to enhance the complex-valued acoustic spectrograms of speech signals via replacing the magnitude part of the corresponding modulation spectrum in order to create noise-robust feature representation for recognition. All the evaluation experiments implemented on the Aurora-2 digit database and task show that the presented method performs better than the baseline MFCC and several well-known noise-robust techniques. These results apparently reveal that this novel method alleviates the effect of noise in speech features significantly.
{"title":"Magnitude replacement of real and imaginary modulation spectrum of acoustic spectrograms for noise-robust speech recognition","authors":"Hsin-Ju Hsieh, J. Hung","doi":"10.1109/ICCE-TW.2015.7216925","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216925","url":null,"abstract":"In this paper, a novel method is proposed to enhance the complex-valued acoustic spectrograms of speech signals via replacing the magnitude part of the corresponding modulation spectrum in order to create noise-robust feature representation for recognition. All the evaluation experiments implemented on the Aurora-2 digit database and task show that the presented method performs better than the baseline MFCC and several well-known noise-robust techniques. These results apparently reveal that this novel method alleviates the effect of noise in speech features significantly.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133251421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-06-06DOI: 10.1109/ICCE-TW.2015.7216839
N. Xu, Yibin Tang, J. Bao, Xiao Yao, A. Jiang, Xiaofeng Liu
In this paper, a computationally efficient voice conversion system has been designed in order to improve the performance in resource-limited scenarios. First, mixtures of Gaussians (MoGs) at fixed locations of Mel frequencies have been used to represent the spectrum of STRAIGHT compactly. Second, the key conditional distributions for prediction are approximated by building histograms of aligned features empirically. Experiments have confirmed that our proposed method can obtain fairly good results compared to the traditional method without huge computational costs.
{"title":"Voice conversion based on empirical conditional distribution in resource-limited scenarios","authors":"N. Xu, Yibin Tang, J. Bao, Xiao Yao, A. Jiang, Xiaofeng Liu","doi":"10.1109/ICCE-TW.2015.7216839","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2015.7216839","url":null,"abstract":"In this paper, a computationally efficient voice conversion system has been designed in order to improve the performance in resource-limited scenarios. First, mixtures of Gaussians (MoGs) at fixed locations of Mel frequencies have been used to represent the spectrum of STRAIGHT compactly. Second, the key conditional distributions for prediction are approximated by building histograms of aligned features empirically. Experiments have confirmed that our proposed method can obtain fairly good results compared to the traditional method without huge computational costs.","PeriodicalId":340402,"journal":{"name":"2015 IEEE International Conference on Consumer Electronics - Taiwan","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127812025","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：