Foreword

Preface

1Introduction

1.1ArtificialAutonomousSystems

1.2NeuralComputationandAnalogIntegratedCircuits

2VisualMotionPerception

2.1ImageBrightness

2.2CorrespondenceProblem

2.3OpticalFlow

2.4MatchingModels

2.5FlowModels

2.6OutlineforaVisualMotionPerceptionSystem

2.7ReviewofaVLSIImplementations

3OptimizationNetworks

3.1AssociativeMemoryandOptimization

3.2ConstraintSatisfactionProblems

3.3Winner-takes-allNetworks

3.4ResistiveNetwork

4VisualMotionPerceptionNetworks

4.1ModelforOpticalFlowEstimation

4.2NetworkArchitecture

4.3SimulationResultsforNaturalImageSequences

4.4PassiveNon-linearNetworkConducatances

4.5ExtendedRecurrentNetworkArchitectures

4.6Remarks

5AnalogVLSIImplementation

5.1ImplementationSubstrate

5.2Phototransduction

5.3ExtractionoftheSpatio-temporalBrightnessGradients

5.4SingleOpticalFlowUnit

5.5Layout

6SmoothOpticalFlowChip

6.1ResponseCharacteristics

6.2Intersection-of-constraintsSolution

6.3FlowFieldEstimation

6.4DeviceMismatch

6.5ProcessingSpeed

6.6Applications

7ExtendedNetworkImplementations

7.1MotionSegmentationChip

7.2MotionSelectionChip

8ComparisontoHumanMotionVision

8.1Humanvs.ChipPerception

8.2ComputationalArchitecture

8.3Remarks

AVariationalCalculus

BSimulationMethods

CTransistorsandBasicCircuits

DProcessParametersandChipsSpecifications

References

Index2100433B