Computational Models for Predicting Visual Target Distinctness

by Jose A. Garcia, R. Rodriguez-Sanchez, Xose R. Fdez-Vidal, Joaquin Fdez-Valdivia, Rosa Rodriguez-Sanchez

Overview

The more a target stands out from its background, the easier it is to detect and the quicker it will be found. This book looks at two situations for predicting visual target distinctness by means of a computer vision model.

Product Details

ISBN-13:	9780819439963
Publisher:	SPIE Press
Publication date:	01/01/2001
Series:	SPIE Press Monograph Series
Pages:	212
Product dimensions:	7.00(w) x 9.98(h) x 0.57(d)

Table of Contents

	Preface	ix
1	Models of feature perception in distortion measure guidance	1
1.1	Introduction	1
1.2	Computational models for feature detection	3
1.2.1	Image features from Laplacian zero-crossings	3
1.2.2	Image features from phase congruency	3
1.2.3	Image features from active sensors	6
1.2.3.1	A data-driven multisensor scheme	10
1.2.3.2	The activated sensors in the multisensor scheme	11
1.3	Error measure guidance	14
1.4	Experimental results	17
1.4.1	Distinctness of targets and their immediate surroundings	17
1.5	The role of integral features for perceiving image discriminability	18
1.5.1	An original image quality model for predicting the visibility of the difference between a pair of images	19
1.5.1.1	The spatial sensitivity function	21
1.5.2	Applications	22
1.5.2.1	Distinctness of targets and their immediate surroundings	23
1.6	Conclusions	24
2	Computational measures based on space-frequency analysis	25
2.1	Introduction	25
2.2	The multichannel organization of images	27
2.2.1	Overview of approach	27
2.2.2	Clumps of energy in the amplitude spectrum	29
2.2.2.1	A spatial to 2D spatial-frequency transformation	30
2.2.2.2	A data-driven multichannel design	33
2.2.2.2.1	The data-driven selection of bands of orientation	33
2.2.2.2.2	The data-driven selection of radial frequency channels	34
2.2.2.3	The selection of the most activated sensors	37
2.2.3	Bank of log-Gabors filters	37
2.2.4	Activated filters in the bank	40
2.3	Filtered response based distinctness measure	40
2.3.1	Selection of fixation points	41
2.3.2	"Filtered-response" (FR) distinctness measure	41
2.4	Integral features based distinctness measure	44
2.4.1	Preattentive stage	44
2.4.2	Integral-feature (IF) distinctness measure	46
2.5	Experimental results	47
2.5.1	Images, apparatus, subjects, and laboratory viewing conditions	48
2.5.2	Predicting visual target distinctness	49
2.5.2.1	Psychophysical target distinctness	50
2.5.2.2	Search experiment	50
2.5.2.3	Results	51
2.5.2.3.1	Psychophysical target distinctness	56
2.5.2.3.2	Computational target distinctness	56
2.5.2.3.3	Experiment 1	58
2.5.2.3.4	Experiment 2	59
2.5.2.3.5	Experiment 3	59
2.5.2.3.6	Experiment 4	60
2.6	Conclusions	62
3	Defining the notion of visual pattern	67
3.1	Introduction	67
3.2	Material and methods	68
3.2.1	Images	68
3.2.2	The RGFF image representational model	68
3.2.2.1	Selection of strongly responding filters	71
3.2.2.2	Distance between filtered responses	71
3.2.2.2.1	The best definition of integral feature for segregating visual patterns	72
3.2.2.2.2	Congruence in integral features between two filtered responses	73
3.2.2.3	Decomposition of the original reference image into its "visual patterns"	75
3.2.2.3.1	Clustering of activated filters	76
3.2.3	Evaluation function	82
3.2.3.1	Datasets	83
3.2.3.2	Psychophysical target distinctness	84
3.2.3.3	Computational target distinctness metric	91
3.3	Results and discussion	91
3.3.1	Experiment 1	93
3.3.2	Experiment 2	97
3.3.3	Experiment 3	98
3.3.4	Experiment 4	102
3.3.5	Experiment 5	102
3.4	Conclusions	109
4	Information theoretic measures	111
4.1	Introduction	111
4.2	Basic axiomatic characterization	113
4.3	Information conservation constraint	118
4.3.1	Selective information gain	119
4.3.2	Properties of the selective information gain	121
4.4	Significance conservation constraint	123
4.4.1	Compound information gain	124
4.4.2	Properties of the compound gain	125
4.5	Comparative study	128
4.5.1	Images and datasets	129
4.5.2	Psychophysical target distinctness	130
4.5.3	Results and discussion	130
4.5.3.1	Experiment 1	130
4.5.3.2	Experiment 2	138
4.5.3.3	Experiment 3	138
4.5.3.4	Statistical Accuracy	140
4.6	Conclusion	144
	Epilogue	147
A	Comparison with other saliency models	157
B	Integral opponent-colors features	161
B.1	Introduction	161
B.2	Preattentive stage	162
B.2.1	RGB to opponent-color encoding transform	164
B.2.2	2D bank of log-Gabors design	165
B.2.3	Activated filters from the bank	166
B.2.4	Fixation points on each filter response	167
B.3	Integration stage	167
B.3.1	Integral opponent-colors features	167
B.3.2	Target distinctness on each activated filter	174
B.4	Decision stage	174
B.5	Applications	175
B.5.1	Distinctness of targets and their immediate surroundings	175
B.5.2	Distinctness of targets in noisy environments	178
B.6	Conclusions	181
C	Forms of gain and divergence	183
D	Calculating derivatives	185
	Bibliography	187
	Index	201

Editorial Reviews

Spanish computer scientists and physicist consider two different situations in predicting visual target distinctness by means of a computer vision model. First, they assume that the structure of the target-and-background scene and the image without a target can be determined exactly, so that the main problem is how to select relevant information into a limited attentional bottleneck. Second, cases in which the structure of the target and non-target scenes may not be determined exactly, so the structure of the images should be characterized statistically by discrete probability distributions. They report the findings of two experiments to investigate the relationship between the different target distinctness measures and the visual target distinctness measured by human observers. Annotation c. Book News, Inc., Portland, OR (booknews.com)

Booknews