The Experts below are selected from a list of 21825 Experts worldwide ranked by ideXlab platform

Lin Zhong-qin - One of the best experts on this subject based on the ideXlab platform.

  • RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision
    2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA), 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Yuan Gao, Lin Zhong-qin
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

Robert Likamwa - One of the best experts on this subject based on the ideXlab platform.

  • ISCA - RedEye: analog ConvNet image sensor architecture for continuous mobile vision
    ACM SIGARCH Computer Architecture News, 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Julian Gao, Lin Zhong
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

  • RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision
    2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA), 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Yuan Gao, Lin Zhong-qin
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

Yunhui Hou - One of the best experts on this subject based on the ideXlab platform.

  • ISCA - RedEye: analog ConvNet image sensor architecture for continuous mobile vision
    ACM SIGARCH Computer Architecture News, 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Julian Gao, Lin Zhong
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

  • RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision
    2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA), 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Yuan Gao, Lin Zhong-qin
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

Mia Polansky - One of the best experts on this subject based on the ideXlab platform.

  • ISCA - RedEye: analog ConvNet image sensor architecture for continuous mobile vision
    ACM SIGARCH Computer Architecture News, 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Julian Gao, Lin Zhong
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

  • RedEye: Analog ConvNet Image Sensor Architecture for Continuous Mobile Vision
    2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA), 2016
    Co-Authors: Robert Likamwa, Yunhui Hou, Mia Polansky, Yuan Gao, Lin Zhong-qin
    Abstract:

    Continuous mobile vision is limited by the inability to efficiently capture image frames and process vision features. This is largely due to the energy burden of analog readout circuitry, data traffic, and Intensive Computation. To promote efficiency, we shift early vision processing into the analog domain. This results in RedEye, an analog convolutional image sensor that performs layers of a convolutional neural network in the analog domain before quantization. We design RedEye to mitigate analog design complexity, using a modular column-parallel design to promote physical design reuse and algorithmic cyclic reuse. RedEye uses programmable mechanisms to admit noise for tunable energy reduction. Compared to conventional systems, RedEye reports an 85% reduction in sensor energy, 73% reduction in cloudlet-based system energy, and a 45% reduction in Computation-based system energy.

Danhuai Guo - One of the best experts on this subject based on the ideXlab platform.

  • a visualization platform for spatio temporal data a data Intensive Computation framework
    International Conference on Geoinformatics, 2015
    Co-Authors: Danhuai Guo
    Abstract:

    Data visualization, as an intuitive approach to help people realize data and knowledge discovering, has been developed with diverse perspectives and objectives, and they may render different analysis results even with the same application case or dataset treated. With the explosive increase of data volume and data dimension, the performance of most of the existing spatio-temporal information visualization toolkits decreases sharply in capacity and efficiency. In this paper, we present a visual analytics platform in data Intensive Computation environment that supports large-scale spatio-temporal data. By redefining task model, data model, and visual mapping strategies, this platform supports processing and visualizing many kinds of Big Data with spatio-temporal attributes. The processing and visualizing can be done in seconds by distributed storage, data reorganization, distributed query, spatial indices, and segmented fetch, even though it has a terabyte of data. In the experimental implementation, the taxi trajectory dataset with 1TB volume and four typical spatio-temporal queries are used to testify our platform's effectiveness and efficiency.

  • Geoinformatics - A visualization platform for spatio-temporal data: A data Intensive Computation framework
    2015 23rd International Conference on Geoinformatics, 2015
    Co-Authors: Danhuai Guo
    Abstract:

    Data visualization, as an intuitive approach to help people realize data and knowledge discovering, has been developed with diverse perspectives and objectives, and they may render different analysis results even with the same application case or dataset treated. With the explosive increase of data volume and data dimension, the performance of most of the existing spatio-temporal information visualization toolkits decreases sharply in capacity and efficiency. In this paper, we present a visual analytics platform in data Intensive Computation environment that supports large-scale spatio-temporal data. By redefining task model, data model, and visual mapping strategies, this platform supports processing and visualizing many kinds of Big Data with spatio-temporal attributes. The processing and visualizing can be done in seconds by distributed storage, data reorganization, distributed query, spatial indices, and segmented fetch, even though it has a terabyte of data. In the experimental implementation, the taxi trajectory dataset with 1TB volume and four typical spatio-temporal queries are used to testify our platform's effectiveness and efficiency.