Regulation of chromatic induction by neighboring images

Margalith Harrar and Françoise Viénot

Author Affiliations

Margalith Harrar^{1} and Françoise Viénot^{1}

^{1}Centre de Recherches sur la Conservation des Documents Graphiques, Muséum National d’Histoire Naturelle, Centre National de la Recherche Scientifique, 36 rue Geoffroy Saint Hilaire, 75005 Paris, France

We deal with the regulation of chromatic contrast when the induction of a second stimulus (one of five neighboring surrounds) opposes the induction from a first stimulus (one of two remote vivid peripheral fields). Using a hue cancellation judgment, we show that, although every neighboring surround that we used has the same average chromatic content, the resulting color appearance of the target differs between surrounds, and this may be ascribed to the spatiochromatic organization of the surround. So, rather than the chromatic contrast amplitude or the frequential structure of the surround, it is the structure of proximity that matters.

Shigeko Takahashi and Yoshimichi Ejima J. Opt. Soc. Am. 73(2) 190-197 (1983)

References

You do not have subscription access to this journal. Citation lists with outbound citation links are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

You do not have subscription access to this journal. Article level metrics are available to subscribers only. You may subscribe either as an OSA member, or as an authorized user of your institution.

Four factors—observer, periphery, surround, and session and their interactions—were analyzed. A difference is significant when the p value is less than 0.05. Results show significant differences for the factors observer, periphery, and surround and for their interactions. No significant difference is found for the factor session and two of its interactions with the other factors. Significant differences are printed in bold numbers.

Table 2

Comparison of the Averages of the Amount of Induction Obtained with Each Neighboring Scene^{
a
}

Red Peripheral Inducing Field

Green Peripheral Inducing Field

Compared Scenes

ε

p

Compared Scenes

ε

p

Pixelized–Rectangles

0.127

0.450

Pixelized–Natural

0.746

0.228

Rectangles–Natural

4.388

$\mathbf{<}\mathbf{0.001}$

Natural–Squares

1.765

0.039

Natural–Beach

2.490

0.006

Squares–Rectangles

0.572

0.284

Beach–Squares

2.481

0.007

Rectangles–Beach

1.289

0.099

Squares–Black

28.825

$\mathbf{<}\mathbf{0.001}$

Beach–Black

23.982

$\mathbf{<}\mathbf{0.001}$

The average l values of the scenes are compared two by two according to the ranking of the l value along the l axis by using ε statistics with Eq. (1). If ε is equal to or higher than 1.96, the difference is significant according to the risk factor defined for $p=0.05$. Significant differences are printed in bold numbers.

Table 3

Comparison of the Variances of the Amount of Induction Obtained with Each Neighboring Scene^{
a
}

Red Peripheral Inducing Field

Green Peripheral Inducing Field

Compared Scenes

F

p

Compared Scenes

F

p

Pixelized–Natural

0.914

0.244

Pixelized–Natural

0.779

0.027

Natural–Rectangles

0.674

0.001

Natural–Rectangles

0.853

0.109

Rectangles–Squares

0.844

0.094

Rectangles–Squares

0.814

0.056

Squares–Beach

0.822

0.065

Squares–Beach

0.949

0.342

Beach–Black

0.661

0.001

Beach–Black

0.835

0.081

The variances of the l value of each neighboring scene are ranked and compared two by two by using an F statistics with Eq. (2). For a risk factor of 0.05, the higher limit is ${F}_{\mathrm{max}}=1.35$, and the lower limit is ${F}_{\mathrm{min}}=0.74$. If the F statistic result falls outside this interval, the two neighboring scenes are considered significantly different. Significant differences are printed in bold numbers.

Table 4

Comparison of the Averages and of the Variances of the l Value Obtained for Groups of Neighboring Scenes^{
a
}

Compared Scenes

Red Peripheral Inducing Field

Green Peripheral Inducing Field

ϵ

p

F

p

ϵ

p

F

p

Nat, Pix, Rec—Bea, Squ

9.992

$\mathbf{<}\mathbf{0.001}$

0.581

$\mathbf{<}\mathbf{0.001}$

3.273

0.001

0.665

$\mathbf{<}\mathbf{0.001}$

Bea, Squ—Bla

32.590

$\mathbf{<}\mathbf{0.001}$

0.603

$\mathbf{<}\mathbf{0.001}$

28.338

$\mathbf{<}\mathbf{0.001}$

0.813

0.030

We have grouped the results obtained with the scenes Natural, Pixelized, and Rectangles on the one hand and the scenes Beach and Squares on the other hand, according to their averages and variances of l value. The comparison between the class of scenes Natural–Pixelized–Rectangles and the class of scenes Beach–Squares, as the comparison between the class of scenes Beach–Squares and Black, show a significant difference in terms of average and variance. Significant differences are printed in bold numbers.

Tables (4)

Table 1

Four-Way ANOVA Table^{
a
}

Factors and Interactions

df

F

p

1—Observer

4

300.05

$\mathbf{<}\mathbf{0.01}$

2—Periphery

2

122.75

$\mathbf{<}\mathbf{0.01}$

3—Surround

5

7.60

$\mathbf{<}\mathbf{0.01}$

4—Session

5

1.05

0.42

1, 2

8

62.28

$\mathbf{<}\mathbf{0.01}$

1, 3

20

26.32

$\mathbf{<}\mathbf{0.01}$

2, 3

10

39.48

$\mathbf{<}\mathbf{0.01}$

1, 4

20

5.00

$\mathbf{<}\mathbf{0.01}$

2, 4

10

1.29

0.27

3, 4

25

1.59

0.06

1–3

40

15.16

$\mathbf{<}\mathbf{0.01}$

1, 2, 4

40

18.89

$\mathbf{<}\mathbf{0.01}$

1, 3, 4

100

3.80

$\mathbf{<}\mathbf{0.01}$

2–4

50

2.06

$\mathbf{<}\mathbf{0.01}$

1–4

200

3.61

$\mathbf{<}\mathbf{0.01}$

Four factors—observer, periphery, surround, and session and their interactions—were analyzed. A difference is significant when the p value is less than 0.05. Results show significant differences for the factors observer, periphery, and surround and for their interactions. No significant difference is found for the factor session and two of its interactions with the other factors. Significant differences are printed in bold numbers.

Table 2

Comparison of the Averages of the Amount of Induction Obtained with Each Neighboring Scene^{
a
}

Red Peripheral Inducing Field

Green Peripheral Inducing Field

Compared Scenes

ε

p

Compared Scenes

ε

p

Pixelized–Rectangles

0.127

0.450

Pixelized–Natural

0.746

0.228

Rectangles–Natural

4.388

$\mathbf{<}\mathbf{0.001}$

Natural–Squares

1.765

0.039

Natural–Beach

2.490

0.006

Squares–Rectangles

0.572

0.284

Beach–Squares

2.481

0.007

Rectangles–Beach

1.289

0.099

Squares–Black

28.825

$\mathbf{<}\mathbf{0.001}$

Beach–Black

23.982

$\mathbf{<}\mathbf{0.001}$

The average l values of the scenes are compared two by two according to the ranking of the l value along the l axis by using ε statistics with Eq. (1). If ε is equal to or higher than 1.96, the difference is significant according to the risk factor defined for $p=0.05$. Significant differences are printed in bold numbers.

Table 3

Comparison of the Variances of the Amount of Induction Obtained with Each Neighboring Scene^{
a
}

Red Peripheral Inducing Field

Green Peripheral Inducing Field

Compared Scenes

F

p

Compared Scenes

F

p

Pixelized–Natural

0.914

0.244

Pixelized–Natural

0.779

0.027

Natural–Rectangles

0.674

0.001

Natural–Rectangles

0.853

0.109

Rectangles–Squares

0.844

0.094

Rectangles–Squares

0.814

0.056

Squares–Beach

0.822

0.065

Squares–Beach

0.949

0.342

Beach–Black

0.661

0.001

Beach–Black

0.835

0.081

The variances of the l value of each neighboring scene are ranked and compared two by two by using an F statistics with Eq. (2). For a risk factor of 0.05, the higher limit is ${F}_{\mathrm{max}}=1.35$, and the lower limit is ${F}_{\mathrm{min}}=0.74$. If the F statistic result falls outside this interval, the two neighboring scenes are considered significantly different. Significant differences are printed in bold numbers.

Table 4

Comparison of the Averages and of the Variances of the l Value Obtained for Groups of Neighboring Scenes^{
a
}

Compared Scenes

Red Peripheral Inducing Field

Green Peripheral Inducing Field

ϵ

p

F

p

ϵ

p

F

p

Nat, Pix, Rec—Bea, Squ

9.992

$\mathbf{<}\mathbf{0.001}$

0.581

$\mathbf{<}\mathbf{0.001}$

3.273

0.001

0.665

$\mathbf{<}\mathbf{0.001}$

Bea, Squ—Bla

32.590

$\mathbf{<}\mathbf{0.001}$

0.603

$\mathbf{<}\mathbf{0.001}$

28.338

$\mathbf{<}\mathbf{0.001}$

0.813

0.030

We have grouped the results obtained with the scenes Natural, Pixelized, and Rectangles on the one hand and the scenes Beach and Squares on the other hand, according to their averages and variances of l value. The comparison between the class of scenes Natural–Pixelized–Rectangles and the class of scenes Beach–Squares, as the comparison between the class of scenes Beach–Squares and Black, show a significant difference in terms of average and variance. Significant differences are printed in bold numbers.