The Making Of A Neuromorphic Visual System

Rasche C.0387234691

The reader is presented an approach to the construction of a visual system, which is behaviorally, computationally and neurally motivated. The central goal is to characterize the process of visual categorization and to find a suitable representation format that can successfully deal with the structural variability existent within visual categories. It does not define such representations a priori but attempts to show directions on how to gradually work towards them. The book reviews past and existent theories of visual object and shape recognition in the fields of computer vision, neuroscience and psychology. The entire range of computations is discussed, as for example contour extraction in retinal circuits, orientation determination in cortical networks, position and scale independence of shape, as well as the issue of object and shape representation in a neural substrate. Region-based approaches are discussed and are modeled with wave-propagating networks. It is demonstrated how those networks operate on gray-scale images. A completely novel shape recognition architecture is proposed that can recognize simple shapes under various degraded conditions. It is discussed how such networks can be used for constructing basic-level object representations. It is envisioned how those networks can be implemented using the method of neuromorphic engineering, an analog electronic hardware substrate than can run neural computations in real-time and with little power.

---

Table of contents :
Cover……Page 1
Table Of Contents……Page 6
1.2 Speed……Page 14
1.4 Recognition Evolvement……Page 15
1.5 Basic-Level Categorization……Page 16
1.7 Summary……Page 18
2.2 Viewpoint Independence……Page 20
2.3 Representation and Evolvement……Page 22
2.3.1 Identification Systems……Page 23
2.3.2 Part-based Descriptions……Page 24
2.3.4 Scene Recognition……Page 26
2.4 Recapitulation……Page 27
2.5 Refining the Primary Engineering Goal……Page 28
3.1 Hierarchy and Models……Page 30
3.2 Criticism and Variants……Page 33
3.3 Speed……Page 36
3.4 Alternative ‘Codes’……Page 38
3.5 Alternative Shape Recognition……Page 40
3.6 Insight from Cases of Visual Agnosia……Page 42
3.7 Neuronal Level……Page 44
3.8 Recapitulation and Conclusion……Page 48
4.1 The Transistor……Page 50
4.2 A Synaptic Circuit……Page 51
4.3 Dendritic Compartments……Page 52
4.5 A Silicon Cortex……Page 53
4.7 Recapitulation……Page 55
5.1 A Representation with Polygons……Page 58
5.2 A Representation with Polygons and their Context……Page 62
5.3 Recapitulation……Page 64
6.2 Spatial Analysis in the Real Retina……Page 68
6.2.1 Method of Adjustable Thresholds……Page 70
6.3 The Propagation Map……Page 71
6.4.2 Method of Latencies……Page 73
6.5 Recapitulation……Page 77
7.1 The Transform……Page 80
7.2 Architecture……Page 81
7.3 Performance……Page 83
7.5 Fast Waves……Page 87
7.6 Recapitulation……Page 88
8.1.2 Biophysical……Page 90
8.2 Speed Detecting Architectures……Page 92
8.3 Simulation……Page 94
8.4 Biophysical Plausibility……Page 96
8.5 Recapitulation……Page 98
9.1 Integration Perspectives……Page 100
9.2 Position and Size Invariance……Page 102
9.3 Architecture for a Template Approach……Page 105
9.4 Basic-Level Representations……Page 107
9.5 Recapitulation……Page 108
10.1 The Idea of the Contour Propagation Field……Page 110
10.2 Architecture……Page 111
10.3 Testing……Page 113
10.4 Discussion……Page 117
10.5 Learning……Page 120
10.6 Recapitulation……Page 122
11.2 Representation, Evolvement, Gist……Page 124
11.3 Scene Exploration……Page 126
11.4 Engineering……Page 128
11.5 Recapitulation……Page 129
12.1 The Quest for Efficient Representation and Evolvement……Page 130
12.3 Neuroscientific Inspiration……Page 134
12.5 Future Approach……Page 135
Terminology……Page 138
References……Page 142
Keywords……Page 150
Abbreviations……Page 152