The Experts below are selected from a list of 65520 Experts worldwide ranked by ideXlab platform

Victor C M Leung - One of the best experts on this subject based on the ideXlab platform.

  • edge network assisted real time object detection framework for autonomous driving
    IEEE Network, 2021
    Co-Authors: Seung Wook Kim, Victor C M Leung
    Abstract:

    Computer vision tasks such as object detection are crucial for the operations of autonomous vehicles (AVs). Results of many tasks, even those requiring high computational power, can be obtained within a short delay by offloading them to edge clouds. However, although edge clouds are exploited, real-time object detection cannot always be guaranteed due to dynamic channel quality. To mitigate this problem, we propose an edge-network-assisted real-time object detection framework (EODF). In an EODF, AVs extract the region of interest (Rols) of the Captured Image when the channel quality is not sufficiently good for supporting real-time object detection. Then AVs compress the Image data on the basis of the Rols and transmit the compressed one to the edge cloud. In so doing, real-time object detection can be achieved due to the reduced transmission latency. To verify the feasibility of our framework, we evaluate the probability that the results of object detection are not received within the inter-frame duration (i.e., outage probability) and their accuracy. From the evaluation, we demonstrate that the proposed EODF provides the results to AVs in real time and achieves satisfactory accuracy.

  • edge network assisted real time object detection framework for autonomous driving
    arXiv: Networking and Internet Architecture, 2020
    Co-Authors: Seung Wook Kim, Victor C M Leung
    Abstract:

    Autonomous vehicles (AVs) can achieve the desired results within a short duration by offloading tasks even requiring high computational power (e.g., object detection (OD)) to edge clouds. However, although edge clouds are exploited, real-time OD cannot always be guaranteed due to dynamic channel quality. To mitigate this problem, we propose an edge network-assisted real-time OD framework~(EODF). In an EODF, AVs extract the region of interests~(RoIs) of the Captured Image when the channel quality is not sufficiently good for supporting real-time OD. Then, AVs compress the Image data on the basis of the RoIs and transmit the compressed one to the edge cloud. In so doing, real-time OD can be achieved owing to the reduced transmission latency. To verify the feasibility of our framework, we evaluate the probability that the results of OD are not received within the inter-frame duration (i.e., outage probability) and their accuracy. From the evaluation, we demonstrate that the proposed EODF provides the results to AVs in real-time and achieves satisfactory accuracy.

Seung Wook Kim - One of the best experts on this subject based on the ideXlab platform.

  • edge network assisted real time object detection framework for autonomous driving
    IEEE Network, 2021
    Co-Authors: Seung Wook Kim, Victor C M Leung
    Abstract:

    Computer vision tasks such as object detection are crucial for the operations of autonomous vehicles (AVs). Results of many tasks, even those requiring high computational power, can be obtained within a short delay by offloading them to edge clouds. However, although edge clouds are exploited, real-time object detection cannot always be guaranteed due to dynamic channel quality. To mitigate this problem, we propose an edge-network-assisted real-time object detection framework (EODF). In an EODF, AVs extract the region of interest (Rols) of the Captured Image when the channel quality is not sufficiently good for supporting real-time object detection. Then AVs compress the Image data on the basis of the Rols and transmit the compressed one to the edge cloud. In so doing, real-time object detection can be achieved due to the reduced transmission latency. To verify the feasibility of our framework, we evaluate the probability that the results of object detection are not received within the inter-frame duration (i.e., outage probability) and their accuracy. From the evaluation, we demonstrate that the proposed EODF provides the results to AVs in real time and achieves satisfactory accuracy.

  • edge network assisted real time object detection framework for autonomous driving
    arXiv: Networking and Internet Architecture, 2020
    Co-Authors: Seung Wook Kim, Victor C M Leung
    Abstract:

    Autonomous vehicles (AVs) can achieve the desired results within a short duration by offloading tasks even requiring high computational power (e.g., object detection (OD)) to edge clouds. However, although edge clouds are exploited, real-time OD cannot always be guaranteed due to dynamic channel quality. To mitigate this problem, we propose an edge network-assisted real-time OD framework~(EODF). In an EODF, AVs extract the region of interests~(RoIs) of the Captured Image when the channel quality is not sufficiently good for supporting real-time OD. Then, AVs compress the Image data on the basis of the RoIs and transmit the compressed one to the edge cloud. In so doing, real-time OD can be achieved owing to the reduced transmission latency. To verify the feasibility of our framework, we evaluate the probability that the results of OD are not received within the inter-frame duration (i.e., outage probability) and their accuracy. From the evaluation, we demonstrate that the proposed EODF provides the results to AVs in real-time and achieves satisfactory accuracy.

Euisik Yoon - One of the best experts on this subject based on the ideXlab platform.

  • A 3.4-w object-adaptive cmos Image sensor with embedded feature extraction algorithm for motion-triggered object-of-interest imaging
    IEEE Journal of Solid-State Circuits, 2014
    Co-Authors: Jaehyuk Choi, Seokjun Park, Jihyun Cho, Euisik Yoon
    Abstract:

    We report a low-power object-adaptive CMOS Imager, which suppresses spatial temporal bandwidth. The object-adaptive Imager has embedded a feature extraction algorithm for identifying objects of interest. The sensor wakes up triggered by motion sensing and extracts features from the Captured Image for the detection of object-of-interest (OOI). Full-Image capturing operation and Image signal transmission are performed only when the interested objects are found, which significantly reduces power consumption at the sensor node. This motion-triggered OOI imaging significantly saves a spatial bandwidth more than 96.5% from the feature output and saves a temporal bandwidth from the motion-triggered wakeup and object adaptive imaging. The sensor consumes low power by employing a reconfigurable differential-pixel architecture with reduced power supply voltage and by implementing the feature extraction algorithm with mixed-signal circuitry in a small area. The chip operates at 0.22 ?W/frame in motion-sensing mode and at 3.4 ?W/frame for feature extraction, respectively. The object detection from on-chip feature extraction circuits has demonstrated a 94.5% detection rate for human from a set of 200 sample Images.

  • a cmos fingerprint system on a chip with adaptable pixel networks and column parallel processors for Image enhancement and recognition
    IEEE Journal of Solid-state Circuits, 2008
    Co-Authors: Seongjin Kim, Kwanghyun Lee, Sangwook Han, Euisik Yoon
    Abstract:

    We propose a 200times160 pixel CMOS fingerprint system-on-a-chip with a local adaptive pixel scheme and embedded column-parallel processors for performing 2-D digital Image processing for fingerprint recognition. The pixel includes a sensing block, ADC, and in-pixel frame memory with no additional area penalty. The sensor can capture robust fingerprint Images in various finger conditions using a locally adapted capacitive sensing scheme through adaptable pixel networks. The embedded parallel processors can enhance the Captured Image in digital signal domain by self-reconfigurable signal processing including low-pass and band-pass filtering, and create the thinned Image for fingerprint recognition. A test chip has been fabricated using a 0.5 mum standard CMOS process. We have successfully Captured and characterized the fingerprint Images from the fabricated test chip at each step of adaptive signal processing steps. The total execution time for acquiring and processing a fingerprint Image is less than 360 ms at 10 MHz and the power consumption is below 70 mW at 3.3 V supply voltage. The proposed sensor can be applied to personal identification systems for portable devices such as cellular phones, PDA and smart cards.

Chi-wah Wong - One of the best experts on this subject based on the ideXlab platform.

  • automatic white balancing using luminance component and standard deviation of rgb components Image preprocessing
    International Conference on Acoustics Speech and Signal Processing, 2004
    Co-Authors: Hong-kwai Lam, Chi-wah Wong
    Abstract:

    Automatic white balancing is an essential Image preprocessing component in consumer digital still cameras, and it can greatly improve the final Image quality of the Captured Image. In this paper, a novel automatic white balancing algorithm based on both the luminance component and standard deviation of RGB components of the pre-Captured Image is proposed. A light source model for evaluation of an automatic white balancing is also described. The simulation results indicate that the proposed algorithm can improve the final Image quality of the Captured Image.

  • Automatic white balancing using luminance component and standard deviation of RGB components
    2004
    Co-Authors: Hong-kwai Lam, Chi-wah Wong
    Abstract:

    Automatic white balancing is an essential Image pre-processing component in consumer digital still cameras, and it can greatly improve the final Image quality of the Captured Image. In this paper, a novel automatic white balancing algorithm based on both luminance component and standard deviation of RGB components of the preCaptured Image is proposed. A light source model for evaluation of an automatic white balancing is also described. The simulation results indicate that the proposed algorithm can improve the final Image quality of the Captured Image.

Masanobu Shinozuka - One of the best experts on this subject based on the ideXlab platform.

  • a vision based system for remote sensing of bridge displacement
    Ndt & E International, 2006
    Co-Authors: Masanobu Shinozuka
    Abstract:

    This study proposed the vision-based system which remotely measures dynamic displacement of bridges in real-time using digital Image processing techniques. This system has a number of innovative features including a high resolution in dynamic measurement, remote sensing, cost-effectiveness, real-time measurement and visualization, ease of installation and operation and no electro-magnetic interference. The digital video camera combined with a telescopic device takes a motion picture of the target installed on a measurement location. Meanwhile, the displacement of the target is calculated using an Image processing technique, which requires a target recognition algorithm, projection of the Captured Image, and calculation of the actual displacement using target geometry and number of pixels moved. For the purpose of verification, a laboratory test using shaking table test and field application on a bridge with open-box girders were carried out. The test results gave sufficient dynamic resolution in frequency as well as the amplitude.

  • real time displacement measurement of a flexible bridge using digital Image processing techniques
    Experimental Mechanics, 2006
    Co-Authors: Masanobu Shinozuka
    Abstract:

    In this study, real-time displacement measurement of bridges was carried out by means of digital Image processing techniques. This is innovative, highly cost-effective and easy to implement, and yet maintains the advantages of dynamic measurement and high resolution. First, the measurement point is marked with a target panel of known geometry. A commercial digital video camera with a telescopic lens is installed on a fixed point away from the bridge (e.g., on the coast) or on a pier (abutment), which can be regarded as a fixed point. Then, the video camera takes a motion picture of the target. Meanwhile, the motion of the target is calculated using Image processing techniques, which require a texture recognition algorithm, projection of the Captured Image, and calculation of the actual displacement using target geometry and the number of pixels moved. Field tests were carried out for the verification of the present method. The test results gave sufficient dynamic resolution in amplitude as well as the frequency. Use of this technology for a large suspension bridge is discussed considering the characteristics of such bridges having low natural frequencies within 3 Hz and the maximum displacement of several centimeters.