Pub Date : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7520975
Cheng-Hung Lin, Wen-Jui Chou
Omni-directional cameras are widely used in many applications such as surveillance systems and endoscopy. Omnidirectional cameras use a single camera and a reflective mirror to capture elliptic omnidirectional images and then transform the elliptic omnidirectional images to panoramic images. To accelerate the transformation from elliptic omnidirectional images to panoramic images, this paper proposes a hierarchical parallelism including data parallelism and task parallelism to improve the performance of transformation using graphic processing units. The data parallelism accelerates the mapping of pixels from elliptic omnidirectional images to panoramic images using multiple threads simultaneously while the task parallelism performs deep pipelines on multiple streams. We have implemented the proposed algorithm using CUDA on NVIDIA GPUs. The experimental results show that the proposed hierarchical parallelism performed on GPUs achieves 6.33 times faster than the CPU counterpart does.
{"title":"Acceleration of the transformation from elliptic omnidirectional images to panoramic images using graphic processing units","authors":"Cheng-Hung Lin, Wen-Jui Chou","doi":"10.1109/ICCE-TW.2016.7520975","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7520975","url":null,"abstract":"Omni-directional cameras are widely used in many applications such as surveillance systems and endoscopy. Omnidirectional cameras use a single camera and a reflective mirror to capture elliptic omnidirectional images and then transform the elliptic omnidirectional images to panoramic images. To accelerate the transformation from elliptic omnidirectional images to panoramic images, this paper proposes a hierarchical parallelism including data parallelism and task parallelism to improve the performance of transformation using graphic processing units. The data parallelism accelerates the mapping of pixels from elliptic omnidirectional images to panoramic images using multiple threads simultaneously while the task parallelism performs deep pipelines on multiple streams. We have implemented the proposed algorithm using CUDA on NVIDIA GPUs. The experimental results show that the proposed hierarchical parallelism performed on GPUs achieves 6.33 times faster than the CPU counterpart does.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"5 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83119724","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7521027
K. Kojima, Hiroki Taniue, J. Kaneko
This paper describes a method to estimate road condition using our developed network-connected manual wheelchair. We have been developing the wheelchair on which torque sensors, an accelerometer and a GPS receiver are implemented, for gathering the road condition data onto our server PC. Our final purpose is to develop a system which display traffic disturbances for manual wheelchairs on the digital map automatically. For this purpose, this study aims to associate the sensor values with road conditions using Mahalanobis distance. In this paper, firstly, our developed wheelchair is explained briefly. Then, characteristics of acquired data is shown. After that, definition of unit space for this problem and calculation of Maharanobis distance are described. Finally, possibility of categorizing road conditions using the Maharanobis distance defined by significance level is explained in detail with the experimental data.
{"title":"Mahalanobis distance-based road condition estimation method using network-connected manual wheelchair","authors":"K. Kojima, Hiroki Taniue, J. Kaneko","doi":"10.1109/ICCE-TW.2016.7521027","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7521027","url":null,"abstract":"This paper describes a method to estimate road condition using our developed network-connected manual wheelchair. We have been developing the wheelchair on which torque sensors, an accelerometer and a GPS receiver are implemented, for gathering the road condition data onto our server PC. Our final purpose is to develop a system which display traffic disturbances for manual wheelchairs on the digital map automatically. For this purpose, this study aims to associate the sensor values with road conditions using Mahalanobis distance. In this paper, firstly, our developed wheelchair is explained briefly. Then, characteristics of acquired data is shown. After that, definition of unit space for this problem and calculation of Maharanobis distance are described. Finally, possibility of categorizing road conditions using the Maharanobis distance defined by significance level is explained in detail with the experimental data.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"7 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80964112","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7521026
N. Funabiki, Yuki Aoyagi, M. Kuribayashi, Wen-Chun Kao
The User-PC computing system (UPC) has been studied to provide a parallel computing platform for members in a group using idling computing resources in personal computers (PCs) of them. UPC adopts the master-worker model where we have implemented the programs for the master on Linux and for the worker on Linux and Windows. However, the current job scheduling method does not consider the real performance of a worker PC. In this paper, we implement a function to measure the performance of a worker PC using two benchmark programs. The experiment results for six PCs in our group show that there are three times difference in the CPU performance and eight times difference in the disk performance.
{"title":"Worker PC performance measurements using benchmarks for user-PC computing system","authors":"N. Funabiki, Yuki Aoyagi, M. Kuribayashi, Wen-Chun Kao","doi":"10.1109/ICCE-TW.2016.7521026","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7521026","url":null,"abstract":"The User-PC computing system (UPC) has been studied to provide a parallel computing platform for members in a group using idling computing resources in personal computers (PCs) of them. UPC adopts the master-worker model where we have implemented the programs for the master on Linux and for the worker on Linux and Windows. However, the current job scheduling method does not consider the real performance of a worker PC. In this paper, we implement a function to measure the performance of a worker PC using two benchmark programs. The experiment results for six PCs in our group show that there are three times difference in the CPU performance and eight times difference in the disk performance.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"10 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89629310","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7520974
Shintaro Hamanaka, S. Kurihara, S. Fukuda, M. Oguchi, Saneyasu Yamaguchi
Android operating system has a function, called LowMemoryKiller, which forcibly terminates application processes when size of available memory is less than the threshold. On reusing the same application again, re-creation of a process is required and takes longer time. ART (Android Runtime environment) has several GC (Garbage Collection) implementations, and choice of GC has effect on size of processes and behavior of LowMemoryKiller. In this paper, we investigate performance of GC implementations and propose a method for choosing GC implementation depending on application size and state. Then, we show our experimental results and demonstrate that our method reduces the number of process terminations cause by LowMemoryKiller.
{"title":"Application state aware GC selection optimization in Android","authors":"Shintaro Hamanaka, S. Kurihara, S. Fukuda, M. Oguchi, Saneyasu Yamaguchi","doi":"10.1109/ICCE-TW.2016.7520974","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7520974","url":null,"abstract":"Android operating system has a function, called LowMemoryKiller, which forcibly terminates application processes when size of available memory is less than the threshold. On reusing the same application again, re-creation of a process is required and takes longer time. ART (Android Runtime environment) has several GC (Garbage Collection) implementations, and choice of GC has effect on size of processes and behavior of LowMemoryKiller. In this paper, we investigate performance of GC implementations and propose a method for choosing GC implementation depending on application size and state. Then, we show our experimental results and demonstrate that our method reduces the number of process terminations cause by LowMemoryKiller.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"72 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89837101","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7521014
Hong-Son Vu, Jiaxian Guo, Kuan-Hung Chen, Shu-Jui Hsieh, D. Chen
Moving objects recognition plays an important role in camera-only active safety systems and intelligent autonomous vehicles. For these applications, reliable detection performance is required; however, pedestrian detection is challenging due to their divergent dressing and action variety. Besides, real-time detection and recognition performance is also critical. This paper aims to optimize the pedestrian detection and recognition by combining both temporal-domain and spatial-domain methods. Accordingly, we first use Background Subtraction (BS) technique to detect moving objects. Then, we use AdaBoost algorithm to classify the detected moving objects into their categories. Experimental results on our datasets show that the proposed approach can speed up 3.3 times in terms of processing rate, with significantly improved detection performance, i.e., at least 17% detection rate increment and 38% false alarm decrement for daytime out-door applications.
{"title":"A real-time moving objects detection and classification approach for static cameras","authors":"Hong-Son Vu, Jiaxian Guo, Kuan-Hung Chen, Shu-Jui Hsieh, D. Chen","doi":"10.1109/ICCE-TW.2016.7521014","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7521014","url":null,"abstract":"Moving objects recognition plays an important role in camera-only active safety systems and intelligent autonomous vehicles. For these applications, reliable detection performance is required; however, pedestrian detection is challenging due to their divergent dressing and action variety. Besides, real-time detection and recognition performance is also critical. This paper aims to optimize the pedestrian detection and recognition by combining both temporal-domain and spatial-domain methods. Accordingly, we first use Background Subtraction (BS) technique to detect moving objects. Then, we use AdaBoost algorithm to classify the detected moving objects into their categories. Experimental results on our datasets show that the proposed approach can speed up 3.3 times in terms of processing rate, with significantly improved detection performance, i.e., at least 17% detection rate increment and 38% false alarm decrement for daytime out-door applications.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"25 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86641632","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7521054
Shih-Ying Chou, Shi-Chiuan Wang, Yu-Cheng Fan
Accompany with touch screen panel and mid-air control have been developed in recent years, people gradually change their usage from tradition keyboard and mouse to the intuitive manner. Mid-air hands operation interface uses RGB-D camera to capture images from space. Than it use captured color and depth information to track hands and gesture to interact with computer.
{"title":"Depth and color-based three dimensional natural user interface","authors":"Shih-Ying Chou, Shi-Chiuan Wang, Yu-Cheng Fan","doi":"10.1109/ICCE-TW.2016.7521054","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7521054","url":null,"abstract":"Accompany with touch screen panel and mid-air control have been developed in recent years, people gradually change their usage from tradition keyboard and mouse to the intuitive manner. Mid-air hands operation interface uses RGB-D camera to capture images from space. Than it use captured color and depth information to track hands and gesture to interact with computer.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"52 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87383931","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7520936
Pengchen Ma, Dan Tao
Incentive mechanism is one of the most critical issues in the field of mobile crowd sensing, and plays an important role on ensuring the quantity of participants and the coverage rate of perceived data. In this paper, we review the existing research, and propose a "5W1H" model to serve the study on incentive mechanism. Moreover, we conduct analysis on how to design incentive mechanism with a case "Noise Map". Finally, we point out some open issues in this research area.
{"title":"“5W1H” model for incentive mechanism in mobile crowd sensing","authors":"Pengchen Ma, Dan Tao","doi":"10.1109/ICCE-TW.2016.7520936","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7520936","url":null,"abstract":"Incentive mechanism is one of the most critical issues in the field of mobile crowd sensing, and plays an important role on ensuring the quantity of participants and the coverage rate of perceived data. In this paper, we review the existing research, and propose a \"5W1H\" model to serve the study on incentive mechanism. Moreover, we conduct analysis on how to design incentive mechanism with a case \"Noise Map\". Finally, we point out some open issues in this research area.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"22 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79906636","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7520898
Tzung-Je Lee, Wu Yu, You-Ting Liu
This paper presents a current sensor for the 16 series Li-ion battery cells. In order to detect the large current of 3.0 A at 57.6 V charging voltage and avoid the gate oxide overdrive problem, the feedback control loop with two source followers and the single stage differential amplifier are used. The proposed design is implemented using a typical 0.25 μm 1P3M 60V BCD process. The sensing current range is from 0 A to 3.0 A. The transimpedance is simulated to be 0.427 V/A.
{"title":"16 Series Li-ion battery cells current sensor","authors":"Tzung-Je Lee, Wu Yu, You-Ting Liu","doi":"10.1109/ICCE-TW.2016.7520898","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7520898","url":null,"abstract":"This paper presents a current sensor for the 16 series Li-ion battery cells. In order to detect the large current of 3.0 A at 57.6 V charging voltage and avoid the gate oxide overdrive problem, the feedback control loop with two source followers and the single stage differential amplifier are used. The proposed design is implemented using a typical 0.25 μm 1P3M 60V BCD process. The sensing current range is from 0 A to 3.0 A. The transimpedance is simulated to be 0.427 V/A.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"6 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79981090","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7521047
Viet-Hang Duong, Yuan-Shan Lee, Bach-Tung Pham, P. Bao, Jia-Ching Wang
In this paper, we introduce a new color image segmentation by using superpixels as feature representation and Manhattan Nonnegative Matrix Factorization (MahNMF) for accurate segmentation. Firstly, the image pixels are grouped into superpixels and considered as the coarse features. The next step is then conducted by factorizing the matrix feature into two nonnegative matrices, which respectively imply representative features and their combination coefficients per superpixel. Exploiting superpixels as features can avoid using too much global information to obtain an advance in time complexity, and using MahNMF can analyze these features for getting segmented image. The experiments show the promise of this new approach.
本文提出了一种新的彩色图像分割方法，利用超像素作为特征表示，利用曼哈顿非负矩阵分解(Manhattan non - negative Matrix Factorization, MahNMF)进行精确分割。首先，将图像像素分组为超像素，并将其作为粗特征;下一步是将矩阵特征分解为两个非负矩阵，这两个非负矩阵分别表示每个超像素的代表性特征及其组合系数。利用超像素作为特征可以避免使用过多的全局信息来获得时间复杂度的提升，使用MahNMF可以分析这些特征来获得分割图像。实验显示了这种新方法的前景。
{"title":"NMF-based image segmentation","authors":"Viet-Hang Duong, Yuan-Shan Lee, Bach-Tung Pham, P. Bao, Jia-Ching Wang","doi":"10.1109/ICCE-TW.2016.7521047","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7521047","url":null,"abstract":"In this paper, we introduce a new color image segmentation by using superpixels as feature representation and Manhattan Nonnegative Matrix Factorization (MahNMF) for accurate segmentation. Firstly, the image pixels are grouped into superpixels and considered as the coarse features. The next step is then conducted by factorizing the matrix feature into two nonnegative matrices, which respectively imply representative features and their combination coefficients per superpixel. Exploiting superpixels as features can avoid using too much global information to obtain an advance in time complexity, and using MahNMF can analyze these features for getting segmented image. The experiments show the promise of this new approach.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"44 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81590042","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 : 2016-05-27DOI: 10.1109/ICCE-TW.2016.7520984
Yu-Shin Wang, Ching-Hwa Cheng
A Hierarchy Multiple-Voltage (HMulti-Vdd) design technique is proposed in this paper which can effectively reduce power consumption. This paper presents an EDA automation design flow that facilitates separation of high-voltage and low-voltage module in synthesis stage. The proposed HMulti-Vdd methodology can be utilized to identify how many voltage domain and how low supplied voltage are better to design a low-power chip, while include the performance estimation. The HMulti-Vdd software tool includes a low-power multi-Vdd chip design optimization process and joint with several commercial circuit synthesis, physical design tools. Using HMulti-Vdd, the designed module voltage assignment is based on power, delay-time and gate-count analysis. HMulti-Vdd can help designer to reduce the Multi-Vdd design manually efforts. For several designed bio-chips have been validates by using this tool, the power consumption can be effectively reduced up to 50%, and the performance loss can be controlled within 5%.
{"title":"A hierarchy multiple-voltage design technique for low-power performance-manageable bio-chips","authors":"Yu-Shin Wang, Ching-Hwa Cheng","doi":"10.1109/ICCE-TW.2016.7520984","DOIUrl":"https://doi.org/10.1109/ICCE-TW.2016.7520984","url":null,"abstract":"A Hierarchy Multiple-Voltage (HMulti-Vdd) design technique is proposed in this paper which can effectively reduce power consumption. This paper presents an EDA automation design flow that facilitates separation of high-voltage and low-voltage module in synthesis stage. The proposed HMulti-Vdd methodology can be utilized to identify how many voltage domain and how low supplied voltage are better to design a low-power chip, while include the performance estimation. The HMulti-Vdd software tool includes a low-power multi-Vdd chip design optimization process and joint with several commercial circuit synthesis, physical design tools. Using HMulti-Vdd, the designed module voltage assignment is based on power, delay-time and gate-count analysis. HMulti-Vdd can help designer to reduce the Multi-Vdd design manually efforts. For several designed bio-chips have been validates by using this tool, the power consumption can be effectively reduced up to 50%, and the performance loss can be controlled within 5%.","PeriodicalId":6620,"journal":{"name":"2016 IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW)","volume":"23 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2016-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88234790","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：