Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Kerneltron: support vector "machine" in silicon.

R Genov1, G Cauwenberghs

  • 1Dept. of Electr. and Comput. Eng., Johns Hopkins Univ., Baltimore, MD, USA.

IEEE Transactions on Neural Networks
|February 5, 2008
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Brain-Silicon Interface for High-Resolution in vitro Neural Recording.

IEEE transactions on biomedical circuits and systems·2013
Same author

VLSI Potentiostat Array With Oversampling Gain Modulation for Wide-Range Neurotransmitter Sensing.

IEEE transactions on biomedical circuits and systems·2013
Same author

Phase-Synchronization Early Epileptic Seizure Detector VLSI Architecture.

IEEE transactions on biomedical circuits and systems·2013
Same author

Wireless micropower instrumentation for multimodal acquisition of electrical and chemical neural activity.

IEEE transactions on biomedical circuits and systems·2013
Same author

OLAM: A wearable, non-contact sensor for continuous heart-rate and activity monitoring.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2012
Same author

Biophysical Neural Spiking, Bursting, and Excitability Dynamics in Reconfigurable Analog VLSI.

IEEE transactions on biomedical circuits and systems·2012
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles

This study introduces the Kerneltron, a novel VLSI processor for real-time object detection in streaming video. It efficiently handles complex Support Vector Machine (SVM) computations, improving generalization and reducing power consumption.

Area of Science:

  • Computer Engineering
  • Machine Learning
  • VLSI Design

Background:

  • Real-time complex object detection in streaming video faces challenges in generalization from sparse data and computational demands.
  • Support Vector Machines (SVMs) offer generalization but require significant computational power for high-dimensional data.

Purpose of the Study:

  • To develop a hardware solution, the Kerneltron, that addresses the computational and generalization challenges in SVM-based object detection.
  • To create a massively parallel architecture for efficient SVM kernel evaluation.

Main Methods:

  • Designed a mixed-signal very large-scale integration (VLSI) processor, the Kerneltron, featuring a fine-grain analog computational array.
  • Implemented a three-transistor unit cell for integrated storage, binary multiplication, and analog accumulation.

Related Experiment Videos

  • Utilized oversampled quantization and bit-serial unary encoding for precise digital output.

    • Main Results:

    • The Kerneltron processor achieves 6.5 GMACS throughput while consuming only 5.9 mW of power.
    • The 3mm x 3mm chip attains 8-bit output resolution.
    • Demonstrated efficient evaluation of kernels over large numbers of vectors in high dimensions.

    Conclusions:

    • The Kerneltron processor effectively supports SVM generalization and offers high computational efficiency for real-time video analysis.
    • This VLSI design provides a power-efficient solution for demanding machine learning tasks in embedded systems.