The Experts below are selected from a list of 5907 Experts worldwide ranked by ideXlab platform

Kazimierz Wiatr - One of the best experts on this subject based on the ideXlab platform.

  • FPL - Dynamic Constant Coefficient Convolvers Implemented in FPGAs
    Lecture Notes in Computer Science, 2002
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    This paper describes different techniques for implementing in FPGAs Convolution Operation for which coefficients are relatively constant. Multiplication is a very basic Operation required for Convolution, therefore a thorough study of different multipliers (Constant Coefficient Multiplier vs. Dynamic Constant Coefficient Multiplier vs. Variable Coefficient Multiplier), their hardware efficiency and dynamic change of the coefficient are described.

  • implementation of Convolution Operation on general purpose processors
    Proceedings 27th EUROMICRO Conference. 2001: A Net Odyssey, 2001
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Convolution (FIR filtering) is a computationally demanding Operation, especially when performed as a two-dimension (2D) Operation in a real time image processing system. Consequently, a great amount of research has been done to perform this Operation more and more efficiently. This paper reviews different architectures of general-purpose processors giving the example of Pentium family. In recent years a rapid development of microprocessors have been observed, however, it can be seen that architecture of microprocessors is already complex and a further rapid increase of the calculation speed rather limited. Dedicated processors implement Convolution Operation very efficiently, nevertheless, the design is not flexible which makes design development, test and upgrade difficult.

  • EUROMICRO - Implementation of Convolution Operation on general purpose processors
    Proceedings 27th EUROMICRO Conference. 2001: A Net Odyssey, 1
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Convolution (FIR filtering) is a computationally demanding Operation, especially when performed as a two-dimension (2D) Operation in a real time image processing system. Consequently, a great amount of research has been done to perform this Operation more and more efficiently. This paper reviews different architectures of general-purpose processors giving the example of Pentium family. In recent years a rapid development of microprocessors have been observed, however, it can be seen that architecture of microprocessors is already complex and a further rapid increase of the calculation speed rather limited. Dedicated processors implement Convolution Operation very efficiently, nevertheless, the design is not flexible which makes design development, test and upgrade difficult.

  • Convolution Operation implemented in FPGA structures for real-time image processing
    ISPA 2001. Proceedings of the 2nd International Symposium on Image and Signal Processing and Analysis. In conjunction with 23rd International Conferen, 1
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Addition is an essential Operation for Convolution (or FIR filters). In FPGAs, addition should be carried out in a standard way employing ripple-carry adders (rather than carry-save adders), which complicates the search for an optimal adder structure as routing order has a substantial influence on the addition cost. Further, complex parameters of addition inputs have been considered e.g. correlation between inputs. These parameters are specified in different ways for different convolver architectures: multiplierless multiplication, look-up table based multiplication, distributed arithmetic. Furthermore, different optimisation techniques, exhausted search and simulated annealing, have been implemented. Otherwise, exhausted search should be employed for the number of the addition inputs n/spl les/8 or simulated annealing for n>8. Employing simulated annealing gives about 10-20% area reduction in comparison to the greedy algorithm. This paper is a part of the research on the AuToCon-automated tool for generating Convolution in FPGAs.

Youngjin Choi - One of the best experts on this subject based on the ideXlab platform.

  • Faster and smoother trajectory generation considering physical system limits under discontinuously assigned target angles
    2012 IEEE International Conference on Mechatronics and Automation, 2012
    Co-Authors: Geon Joon Lee, Doik Kim, Youngjin Choi
    Abstract:

    This paper suggests a trajectory generation method using Convolution Operation. The proposed trajectory method always generates differentiable S-curve shape regardless of irregular time intervals, even when a new target is inputted before the trajectory does not reach to the previous target, or when a target is lost due to the communication problems. Also, the proposed method generates a trajectory considering physical system limits, for instance, maximum velocity, acceleration, and jerk, through the successive digital Convolution that can reduce computational load, by using a recursive form of Convolution Operation. The effectiveness of the proposed method is shown through simulations.

  • CDC-ECE - New robotic motion generation using digital Convolution with physical system limitation
    IEEE Conference on Decision and Control and European Control Conference, 2011
    Co-Authors: Geon Joon Lee, Doik Kim, Youngjin Choi
    Abstract:

    This paper proposes a novel trajectory generation method considering physical system limitations such as maximum velocity, acceleration, and jerk. Also, the trajectory generation method uses a digital Convolution for reducing computational load in robotic motion applications. The method to be proposed has several advantages; first, a continuously differentiable trajectory is simply obtained by applying successive Convolution Operations. Second, a resultant trajectory is always generated satisfying the given physical system limits. Third, the suggested method has low computational load thanks to recursive form of Convolution Operation. The effectiveness of the proposed method is shown through numerical simulations.

  • Novel Trajectory Generation Method using Convolution Operation
    2010
    Co-Authors: Geon Joon Lee, Youngjin Choi
    Abstract:

    A novel trajectory generation method is proposed by making active use of performance specifications such as maximum speed, acceleration and jerk of a given system. The suggested method has an advantage of less computational burden thanks to a recursive form of Convolution sum in discrete-time domain.

Pengju Tian - One of the best experts on this subject based on the ideXlab platform.

  • evaluation of Convolution Operation based on the interpretation of deep learning on 3 d point cloud
    IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020
    Co-Authors: Bufan Zhao, Xianghong Hua, Wuyong Tao, Shaoquan Feng, Pengju Tian
    Abstract:

    The interpretation of deep learning network is an important part in understanding the Convolutional neural networks (CNNs). As an exploratory research, this article explored the interpretation method in 3-D point cloud deep learning networks, for the purpose of evaluating the performance of Convolution functions in 3-D point cloud CNNs. Specifically, a 3-D point cloud classification network with two branches is used as the interpretation network in two aspects; 1) information entropy is introduced to diagnose the internal representation in the middle layer of CNN; and 2) the external consistency of Convolution function is measured by per-point classification accuracy with class activation mapping technique. Four typical Convolution functions are tested by the interpretation network on ModelNet40 dataset and the experimental results demonstrate that the proposed evaluation method is reliable. Feature transformation ability and feature recognition ability of Convolution functions are extracted by visualization evaluation and proposed measurable metrics evaluation.

Ernest Jamro - One of the best experts on this subject based on the ideXlab platform.

  • FPL - Dynamic Constant Coefficient Convolvers Implemented in FPGAs
    Lecture Notes in Computer Science, 2002
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    This paper describes different techniques for implementing in FPGAs Convolution Operation for which coefficients are relatively constant. Multiplication is a very basic Operation required for Convolution, therefore a thorough study of different multipliers (Constant Coefficient Multiplier vs. Dynamic Constant Coefficient Multiplier vs. Variable Coefficient Multiplier), their hardware efficiency and dynamic change of the coefficient are described.

  • implementation of Convolution Operation on general purpose processors
    Proceedings 27th EUROMICRO Conference. 2001: A Net Odyssey, 2001
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Convolution (FIR filtering) is a computationally demanding Operation, especially when performed as a two-dimension (2D) Operation in a real time image processing system. Consequently, a great amount of research has been done to perform this Operation more and more efficiently. This paper reviews different architectures of general-purpose processors giving the example of Pentium family. In recent years a rapid development of microprocessors have been observed, however, it can be seen that architecture of microprocessors is already complex and a further rapid increase of the calculation speed rather limited. Dedicated processors implement Convolution Operation very efficiently, nevertheless, the design is not flexible which makes design development, test and upgrade difficult.

  • EUROMICRO - Implementation of Convolution Operation on general purpose processors
    Proceedings 27th EUROMICRO Conference. 2001: A Net Odyssey, 1
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Convolution (FIR filtering) is a computationally demanding Operation, especially when performed as a two-dimension (2D) Operation in a real time image processing system. Consequently, a great amount of research has been done to perform this Operation more and more efficiently. This paper reviews different architectures of general-purpose processors giving the example of Pentium family. In recent years a rapid development of microprocessors have been observed, however, it can be seen that architecture of microprocessors is already complex and a further rapid increase of the calculation speed rather limited. Dedicated processors implement Convolution Operation very efficiently, nevertheless, the design is not flexible which makes design development, test and upgrade difficult.

  • Convolution Operation implemented in FPGA structures for real-time image processing
    ISPA 2001. Proceedings of the 2nd International Symposium on Image and Signal Processing and Analysis. In conjunction with 23rd International Conferen, 1
    Co-Authors: Ernest Jamro, Kazimierz Wiatr
    Abstract:

    Addition is an essential Operation for Convolution (or FIR filters). In FPGAs, addition should be carried out in a standard way employing ripple-carry adders (rather than carry-save adders), which complicates the search for an optimal adder structure as routing order has a substantial influence on the addition cost. Further, complex parameters of addition inputs have been considered e.g. correlation between inputs. These parameters are specified in different ways for different convolver architectures: multiplierless multiplication, look-up table based multiplication, distributed arithmetic. Furthermore, different optimisation techniques, exhausted search and simulated annealing, have been implemented. Otherwise, exhausted search should be employed for the number of the addition inputs n/spl les/8 or simulated annealing for n>8. Employing simulated annealing gives about 10-20% area reduction in comparison to the greedy algorithm. This paper is a part of the research on the AuToCon-automated tool for generating Convolution in FPGAs.

Geon Joon Lee - One of the best experts on this subject based on the ideXlab platform.

  • Faster and smoother trajectory generation considering physical system limits under discontinuously assigned target angles
    2012 IEEE International Conference on Mechatronics and Automation, 2012
    Co-Authors: Geon Joon Lee, Doik Kim, Youngjin Choi
    Abstract:

    This paper suggests a trajectory generation method using Convolution Operation. The proposed trajectory method always generates differentiable S-curve shape regardless of irregular time intervals, even when a new target is inputted before the trajectory does not reach to the previous target, or when a target is lost due to the communication problems. Also, the proposed method generates a trajectory considering physical system limits, for instance, maximum velocity, acceleration, and jerk, through the successive digital Convolution that can reduce computational load, by using a recursive form of Convolution Operation. The effectiveness of the proposed method is shown through simulations.

  • CDC-ECE - New robotic motion generation using digital Convolution with physical system limitation
    IEEE Conference on Decision and Control and European Control Conference, 2011
    Co-Authors: Geon Joon Lee, Doik Kim, Youngjin Choi
    Abstract:

    This paper proposes a novel trajectory generation method considering physical system limitations such as maximum velocity, acceleration, and jerk. Also, the trajectory generation method uses a digital Convolution for reducing computational load in robotic motion applications. The method to be proposed has several advantages; first, a continuously differentiable trajectory is simply obtained by applying successive Convolution Operations. Second, a resultant trajectory is always generated satisfying the given physical system limits. Third, the suggested method has low computational load thanks to recursive form of Convolution Operation. The effectiveness of the proposed method is shown through numerical simulations.

  • Novel Trajectory Generation Method using Convolution Operation
    2010
    Co-Authors: Geon Joon Lee, Youngjin Choi
    Abstract:

    A novel trajectory generation method is proposed by making active use of performance specifications such as maximum speed, acceleration and jerk of a given system. The suggested method has an advantage of less computational burden thanks to a recursive form of Convolution sum in discrete-time domain.