Open Access
Add to BibSonomy
Abstract
We investigated the diffuseness of illumination to examine which diffuseness condition faithfully reproduces the surface appearance of an object as seen in a natural environment. We also examined the diffuseness condition which produces the ideal appearance of the object. Observers first memorized the appearance of various objects in daily environments, and then evaluated the appearance of the objects under different diffuseness conditions. The observers reported that the moderate diffuseness condition best reproduced a faithful and ideal appearance of the objects, compared with the low and high diffuseness conditions. This indicates that a very low or high diffuseness, which is unfamiliar, is not suitable for reproducing an object’s surface appearance faithfully and ideally. Our results suggest that it is possible to determine a suitable diffuseness condition for reproducing the appearance of objects.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. INTRODUCTION
The impression of an object changes depending on its illumination; this phenomenon is often experienced by us in daily life. For example, a product bought under the lighting of a shop may have a different appearance when viewed outside or at home. In recent years, with the evolution of solid-state lamps, including light-emitting diodes (LEDs) and organic light-emitting diodes, the development of lighting with various types of intensity, spectral power distribution, and spatial distribution has progressed. In addition, lamps in various shapes, such as plate-type and curved designs, have also been developed. Lighting conditions produced by these systems can exhibit a large variety of lighting distributions. In particular, the diffuseness of lighting changed by lighting distributions can have a significant impact on the appearance of objects. For example, Fig. 1 shows the appearance of an object under illumination with different diffuseness values. Since a change in diffuseness can result in such huge differences in the appearance of an object, it is important to determine the diffuseness condition that is most suitable for reproducing the surface appearance of objects.
The appearance of an object largely depends on its material and shape. Studies on the appearance of texture have been conducted using various material and surface conditions [1–3]. However, regarding the color of an object, Maloney et al. [4] showed that stable color perception can be achieved without depending on the roughness or glossiness of the object’s surface.
The effect of surface texture has been extensively studied with images rendered using the bidirectional reflectance distribution function [5–10]. Lighting conditions that affect an object’s appearance are related to how the light hits the object and how the light spreads [11]. Olkkonen and Brainard [12] showed that the impression of matte objects remained stable, but the impression of glossy objects changed when the lighting environment changed. Mizokami et al. [13] suggested that a rough wavy surface with high frequency tends to appear less glossy and smoother as the diffuseness of the light increases. Moreover, a glossy surface appeared less glossy, and a matte surface appeared smoother under diffusive light. Therefore, it is necessary to estimate the lighting conditions that can best reproduce the surface appearance of an object.
Pont et al. [14] showed that observers judged the glossiness of an object incorrectly when the direction and diffuseness of the illumination changed. Yamazoe et al. [15] showed that a high diffuseness of illumination increases the fidelity of the appearance of an object. However, in their experiment, the diffuseness was limited to a range from low to middle. Therefore, it is not yet clear how much an object’s appearance changes due to different diffuseness levels, and which diffuseness levels reproduce the object’s appearance more faithfully and ideally. In this study, we investigated the diffuseness of illumination to reproduce an object’s faithful and ideal surface appearance, using a wide range of diffuseness levels.
In Experiment I, we examined lighting conditions that reproduced the most faithful and ideal appearance of the given objects across four diffuseness levels. In Experiment II, we added a lower diffuseness condition and used a new evaluation method to clarify suitable lighting conditions. In Experiment III, we focused on the material of the objects; we used two objects of the same material with different amounts of luster, as well as two types of metals.
2. EXPERIMENT I
In Experiment I, we investigated lighting conditions that reproduced a faithful and ideal appearance of an object. The observers memorized four types of stimuli made from different materials in a daily environment. Then, they observed the stimuli under four different diffuseness conditions and evaluated their impression, and rated the lighting condition that was faithful to the appearance of the memorized object and the condition that exhibited the ideal reproduction of the object’s appearance.
A. Environment
A viewing booth was constructed by combining a lighting part and an observation part, imitating a pseudo-integrating sphere, as shown in Fig. 2. An observer viewed the stimulus from the observation holes. The experiments were conducted in a dark room. The lighting part consisted of a combination of an LED lamp (Panasonic EVERLEDS LDA8DA1D; correlated color temperature, 6500 K; color rendering index, ${R_a}$ 73) with a duct and a Fresnel lens. The observation part was made of a white Styrofoam sphere with a diameter of 60 cm to create a high diffuseness lighting condition.
The diffuseness of illumination was set at four levels (0.40, 0.55, 0.67, and 0.93) by changing the length of the duct and the distance between the LED lamp and the lighting port, as shown in Fig. 3. We used the cubic illuminance measurement defined by Cuttle [16] and the diffuseness value defined by Xia et al. [17] as the measurement of diffuseness in our experiment. The illuminance in six directions (tilt angle of ${+}{35}^\circ$ with a rotation angle of 0° [$E_{({\rm{u}} +)}$], 120° [$E_{({\rm{v}} +)}$], and 240° [$E_{({\rm{w}} +)}$]; tilt angle of ${-}{35}^\circ$ with a rotation angle of 60° [$E_{({\rm{u}} -)}$], 180° [$E_{({\rm{v}} -)}$], and 300° [$E_{({\rm{w}} -)}$]) was measured. The illumination vector components on the $u$, $v$, and $w$ axes were [Eq. (1)]
The illumination vector magnitude [Eq. (2)]:
The symmetric components on the $u$, $v$, and $w$ axes [Eq. (3)]:
The symmetric illuminance [Eq. (4)]:
The scalar illuminance [Eq. (5)]:
The diffuseness [Eq. (6)]:
The horizontal illuminance at the position of a stimulus was unified to 300 lx.
Diffuseness was measured using an illuminometer T-10 (KONICA MINOLTA) and a measurement holder created by a 3D printer. The measurement holder is shown in Fig. 4. The holder had six holes to place the sensor head of the illuminometer to fix the measurement angle. An overview of the diffuseness measurements is shown in Fig. 5. We placed the holder at the position of the stimulus and measured the illuminance in six directions. Table 1 shows the illuminance for each diffuseness condition.
Table 1. Illuminance for Each Diffuseness Condition
B. Stimuli
The experimental stimuli were a polyresin ball, fur charm, wooden cube, and stainless-steel cube. Figure 6 shows the experimental stimuli under each diffuseness condition. We measured the chromaticity of the stimuli using a CM-700d spectrophotometer (KONICA MINOLTA). Objects with a spherical surface, such as polyresin balls, were measured at the smallest measurement area; thus, the spectrophotometer was closely attached, and the measurement error was minimal. The fur charm was measured with the measurement surface as flat as possible. However, this pattern remained uneven. Therefore, we used the average of the three measurements to compensate unevenness. Table 2 shows the $L^*$, $a^*$, and $b^*$ values for each stimulus. Table 3 lists the measurement conditions used.
Fig. 6. Experimental stimuli in Experiment I: polyresin ball, fur charm, wooden cube, and stainless-steel cube, in order from the top. Diffuseness levels: 0.40 (low), 0.55 (middle), 0.67 (high), and 0.93 (very high), in order from the left.
Table 2. $L^*$, $a^*$, and $b^*$ Values of Each Stimulus from all Three Experiments
Table 3. Measurement Conditions of CM-700d Spectrophotometer (KONICA MINOLTA)
C. Evaluation Methods
The semantic differential scale (SD) method (3-0-3) with 10 items was used to evaluate the following impressions of the objects: “Transparency,” “Brightness,” “Visibility,” “Preference,” “Glossiness,” “Value,” “Roughness,” “Naturalness,” “Hardness,” and “Weight.” These impression evaluation items were selected covering the judgments in terms of physical properties (“Transparency,” “Hardness,” and “Weight”); surface properties (“Glossiness,” “Brightness,” and “Roughness”), which are often used in research on material perception [18,19]; and more subjective impressions by the observers (“Visibility,” “Preference,” “Value,” and “Naturalness”).
Fig. 8. (a) Average selection rate of all observers for fidelity in Experiment I. (b) The average selection rate of all observers for ideality. Error bars indicate the standard deviation.
In addition to the impression evaluation, observers selected the best illumination condition that faithfully reproduced the memorized object’s appearance (fidelity) and the best illumination condition that ideally produced the features of the object’s surface (ideality). Fidelity is the judgement based on the appearance of the object, which is memorized in advance; ideality is the judgment based on the intrinsic property of the object that the observer expects, regardless of its appearance which is memorized in advance.
D. Procedure
In a pre-observation phase, observers memorized the appearance of the stimuli by viewing and touching them in a natural environment: a north window and an office room, for 5 min each with a total of 10 min, on two separate days. At the window, the illuminance ranged from 430 lx to 5340 lx, and the diffuseness ranged from 0.27 to 0.63. The room lighting consisted of white fluorescent lamps (National FHF32EX-NH; CCT 5000 K; ${R_a}$ 84); the illuminance ranged from 100 lx to 500 lx, and the diffuseness ranged from 0.49 to 0.73.
In the evaluation phase, observers viewed the stimuli in the same environment as the pre-observation phase and performed an impression evaluation (pre-evaluation). The observers then moved to the viewing booth and evaluated the object under the experimental conditions (main evaluation).
In the main evaluation, observers viewed one of the stimuli in four different diffuseness conditions in the viewing booth. There was no fixed observation time for the stimulus, and the order of the lighting conditions was randomized. After viewing the object under all the diffuseness conditions, observers performed an appearance evaluation for fidelity and ideality. This procedure was repeated for all stimuli. The evaluation phase consisted of five sessions.
E. Observers
Three men and two women, in their twenties, participated in the experiment.
F. Results
The profile of the average impression evaluation of all observers is shown in Fig. 7. The condition of very high diffuseness tended to change the impression of the polyresin and stainless-steel objects. The average selection rates of the appearance evaluation by all observers are shown in Figs. 8(a) and 8(b). The error bars represent the standard deviation, which indicates the variability among the observers. For fidelity, the selection ratio of “middle diffuseness (0.55)” was the highest. This diffuseness condition was closest to the average diffuseness (0.51) of the pre-observation phase. For ideality selection, the polyresin and stainless-steel stimuli had a high selection rate for the low diffuseness condition, while the fur and wood stimuli had a high selection rate for the high diffuseness condition. One-way ANOVA was used for the analysis. The selection of the best diffuseness condition for fidelity showed significant differences for the fur charm (${\rm{F}}[{{3}},{{12}}] = {7.310}$, $p = {0.005}$), wooden cube (${\rm{F}}[{{3}},{{12}}] = {3.692}$, $p = {0.043}$), and stainless-steel cube (${\rm{F}}[{{3}},{{12}}] = {3.216}$, $p = {0.062}$). Further, the low diffuseness condition obtained a high rating for fidelity in the case of the polyresin ball, whereas the middle diffuseness condition received a high fidelity rating for the other stimuli. The results of the multiple comparison test using Bonferroni’s method only showed a significant difference between the low and middle diffuseness conditions for the fur charm. Ideality selection did not show any significant differences.
The correlation coefficients between the impression ratings and the selection ratings of fidelity and ideality are shown in Table 4. The maximum correlation coefficient was 0.48 for “Ideality” and “Visibility.” It is possible that the object appearance is ideally reproduced under lighting, which makes the object appear clear. “Naturalness” was moderately correlated with both fidelity and ideality, suggesting that natural impression is an important factor in appearance reproduction.
Table 4. Correlations between Impression Ratings, the Selection Rating of Fidelity, and Ideality
The range of diffuseness in the present experiment does not cover low diffuseness. We needed a more extensive diffuseness range to investigate the comprehensive relationship between the appearance of objects and diffuseness of illumination. It is possible that we did not find any significant differences in most of the conditions in this experiment because the evaluation method was not sufficient to clarify these differences. Therefore, we conducted Experiment II, using a wider range of diffuseness and different evaluation method.
3. EXPERIMENT II
In Experiment II, we investigated lighting conditions that reproduced a faithful and ideal appearance of objects under a wider range of diffuseness conditions. Observers memorized the same stimuli as in Experiment I in a daily environment. Then, they observed the stimuli under five different diffuseness conditions and rated the lighting condition that best reproduced a faithful and ideal appearance of the memorized object. We used a 10-point scaling method for the appearance evaluation. Impression evaluation was not performed in this experiment.
A. Environment
The viewing booth was the same as in Experiment I, except that we added a low diffuseness condition by covering the inside of the upper half of the Styrofoam sphere with a black cloth to extend the diffuseness range. We tested five diffuseness conditions (0.26, 0.40, 0.55, 0.67, and 0.93) in Experiment II, as shown in Fig. 9.
B. Stimuli
We used the same objects as in Experiment I as the stimuli. Figure 10 shows the appearance of each stimulus under each diffuseness condition.
Fig. 10. Stimuli in Experiment II. From top to bottom: polyresin ball, fur charm, wooden cube, and stainless-steel cube. From left to right: diffuseness level 0.26 (very low), 0.40 (low), 0.55 (middle), 0.67 (high), and 0.93 (very high).
C. Evaluation Method
Under each diffuseness condition, observers evaluated the extent to which the stimuli appeared faithful to the objects memorized in the daily environment, and the extent to which the stimuli ideally represented the objects; all ratings were performed on a 10-point scale (1–10).
Fig. 11. (a) Average fidelity score for all observers in Experiment II. (b) The average ideality score for all observers. Error bars indicate the standard deviation.
D. Procedure
A pre-observation was performed following the same procedure as in Experiment I. At the window, the illuminance ranged from 50 lx to 4650 lx, and the diffuseness ranged from 0.33 to 0.63. In the room, the illuminance ranged from 150 lx to 450 lx, and the diffuseness ranged from 0.45 to 0.69.
In the main evaluation, the observer viewed each stimulus in the viewing booth under five lighting conditions, and then performed an appearance evaluation. There was no time limit for the observation, and the order of the illumination conditions was randomized. The stimuli were also presented in a random order, and all four types of stimuli were evaluated in one session. In total, five sessions were conducted.
E. Observers
The observers included two men who had participated in Experiment I and two women who had not participated in Experiment I.
F. Results
Figures 11(a) and 11(b) show the average score of the appearance evaluation for all observers for each stimulus. The error bars represent the standard deviation, indicating the variability among the observers. The middle and high diffuseness conditions were rated highly in terms of fidelity and ideality of the illuminated objects. In the score for fidelity, there were significant differences for the polyresin (${\rm{F}}[{{4}},{{12}}] = {6.638}$, $p = {0.005}$) and stainless-steel stimuli (${\rm{F}}[{{4}},{{12}}] = {5.428}$, $p = {0.010}$). In the evaluation for ideality, there were significant differences for the polyresin (${\rm{F}}[{{4}},{{12}}] = {4.484}$, $p = {0.019}$), wood (${\rm{F}}[{{4}},{{12}}] = {2.920}$, $p = {0.067}$), and stainless-steel stimuli (${\rm{F}}[{{4}},{{12}}] = {6.416}$, $p = {0.005}$). The low diffuseness condition obtained a high score in the ideality evaluation for the stainless-steel cube and a medium score for the other objects.
In Experiments I and II, we found that the appearance of objects differed greatly depending on the presence or absence of a glossy surface. The polyresin and stainless-steel objects that had glossy surfaces showed a high diffuseness dependency, while the wooden and fur objects with matte surfaces showed a low dependency. To investigate the effect of the change in glossiness in more detail, we conducted Experiment III using the same material for the stimuli, but with different amounts of surface gloss, and stimuli with a metallic surface.
4. EXPERIMENT III
In Experiment III, we focused on surface glossiness and conducted an experiment using stimuli made from the same material, but with different degrees of gloss, and stimuli with a metallic gloss. Observers first memorized the stimuli in a daily environment, and then observed them under five different diffuseness conditions and evaluated the lighting condition that best produced a faithful or an ideal appearance of the memorized object. We used the SD method (3-0-3) for impression and appearance evaluation.
A. Environment
The viewing booth and diffuseness conditions were the same as in Experiment II.
B. Stimuli
The experimental stimuli were a glossy polyresin ball, a matte polyresin ball, a stainless-steel cube, and a gold hemisphere. Figure 12 shows the experimental stimuli placed under each diffuseness condition.
Fig. 12. Stimuli in Experiment III. From top to bottom: glossy polyresin ball, matte polyresin ball, stainless-steel cube, and gold hemisphere. From left to right: diffuseness level 0.26 (very low), 0.40 (low), 0.55 (middle), 0.67 (high), and 0.93 (very high).
C. Evaluation Methods
The SD method (3-0-3) with 8 items was used to evaluate the following impressions of the objects: “Clarity,” “Visibility,” “Preference,” “Glossiness,” “Value,” “Roughness,” “Naturalness” and “Weight.” In addition to the impression evaluation, observers evaluated the fidelity and ideality of the illumination conditions for each object. Observers rated the appearance of objects as faithful to the memorized one or as an ideal representation of the memorized object under different lighting conditions, using the same (3-0-3) scale. Interpolation of 0.5 was allowed for the evaluation.
D. Procedure
In a pre-observation phase, the observers memorized the appearance of the stimuli by viewing and touching them at a window and in an office room, for 3 min each with a total of 6 min, on two separate days. At the window, the illuminance ranged from 1250 lx to 6210 lx, and the diffuseness ranged from 0.38 to 0.68. In the room, the illuminance ranged from 130 lx to 460 lx, and the diffuseness ranged from 0.56 to 0.74.
In the evaluation phase, the observers viewed the stimuli in the same environment as in the pre-observation phase, and then performed an impression evaluation in the room as a pre-evaluation. In the main evaluation, the observers viewed each stimulus under five different lighting conditions in the viewing booth and performed an impression and appearance evaluation. There was no time limit for observation, and the order of the lighting conditions was randomized. The stimuli were also presented in a random order, and all four types of stimuli were evaluated in one session. A total of five sessions were conducted.
E. Observers
The observers were two men who had participated in Experiments I and II, and two women who had participated in Experiment I.
F. Results
The profile of the average impression evaluation of all observers is shown in Fig. 13. The condition of very high diffuseness tended to change the impression of all the stimuli. Figures 14(a) and 14(b) show the average scores of fidelity and ideality for all observers and all stimuli. The error bars represent the standard deviation, which indicates the variability among the observers. In the fidelity selection, there were significant differences for the glossy polyresin (${\rm{F}}[{{4}},{{12}}] = {18.820}$, $p = {0.000}$), matte polyresin (${\rm{F}}[{{4}},{{12}}] = {7.664}$, $p = {0.003}$), stainless-steel (${\rm{F}}[{{4}},{{12}}] = {28.616}$, $p = {0.000}$), and gold stimuli (${\rm{F}}[{{4}},{{12}}] = {17.418}$, $p = {0.000}$). In the ideality selection, there were significant differences for the glossy polyresin (${\rm{F}}[{{4}},{{12}}] = {10.142}$, $p = {0.001}$), matte polyresin (${\rm{F}}[{{4}},{{12}}] = {6.598}$, $p = {0.005}),$ stainless-steel (${\rm{F}}[{{4}},{{12}}] = {18.596}$, $p = {0.000}$), and gold stimuli (${\rm{F}}[{{4}},{{12}}] = {12.139}$, $p = {0.000}$).
Fig. 14. (a) Average fidelity score of all observers in Experiment III. (b) The ideality average score of all observers. Error bars indicate the standard deviation.
As in Experiment II, the medium diffuseness condition had a high score. Compared to Experiment II, the significant differences were more pronounced, and a clear trend was observed. The range of diffuseness during the pre-observation period ranged from 0.38 to 0.74. The evaluation of the diffuseness conditions included in the range was high for both fidelity and ideality of illumination.
The correlation coefficients between the impression ratings, fidelity scores, and ideality scores are shown in Table 5. The ratings of “Naturalness,” “Glossiness,” and “Clarity” are moderately correlated with the fidelity and ideality scores. This suggests that these impression items are closely related to the judgment of fidelity and ideality.
Table 5. Correlations between Impression Ratings, Fidelity Scores, and Ideality Scores
5. DISCUSSION
In Experiment I, the diffuseness suitable for faithful and ideal reproduction of the object’s appearance differed depending on the object’s material. In Experiment II, the middle diffuseness condition was found most suitable in terms of a faithful and ideal reproduction of the appearance for all objects. In Experiment III, the middle diffuseness condition was highly rated in terms of fidelity and ideality, even for objects of the same material with different amounts of surface gloss.
It should be noted that we conducted experiments with a limited number of observers and lighting conditions. Although each of the three experiments had four to five observers, all observers scored the highest for moderate diffuseness conditions throughout the experiments, and there were significant differences in all cases in Experiment III. Therefore, we consider that our results are reliable enough to show the trend that the middle diffuseness condition is most suitable for faithful and ideal reproduction of the object’s appearance. However, in order to establish a generalized model for the diffuseness of illumination suitable for reproducing a faithful and ideal appearance, it would be necessary to examine a larger number of observers and a greater variety of lighting environments.
In the experiment by Yamazoe et al. [15], the range of diffuseness was 0.18 to 0.43, and the fidelity increased with an increase in diffuseness. We conducted experiments using a wider range of diffuseness, from 0.26 to 0.91. The “very low” (0.26) and “low” (0.40) diffuseness conditions were within the range tested by Yamazoe et al., and similar results were obtained for all glossy objects.
To include and discuss the fur and wooden objects used in Experiment II, the results of Experiment II were converted to the scale of the results of Experiment III using the equation [Eq. (7)]. $X$ is the score of Experiment II, and $Y$ is the score after scale transformation:
Figure 15 shows the fidelity scores for all experimental stimuli. The middle diffuseness condition had a high fidelity score for all objects. The range of diffuseness at pre-observation in Experiment II was 0.33 to 0.69, and that in Experiment III was 0.38 to 0.74; the low (0.40), middle (0.55), and high (0.67) diffuseness conditions were included in this range. Therefore, the reason for the high rating of the middle diffuseness condition could be that the observers found this condition to closely match the diffuseness level of the pre-observation environment. Another reason could be that the observers evaluated the diffuseness levels based on what they experience in daily life [17,20]. Xia et al. [17] found that 0.5–0.7 is the natural diffuseness range. The conditions of 0.56 and 0.67, respectively, which were highly evaluated in our experiment, are included in that range, indicating that our results are consistent with their outcomes. Fleming et al. [21] showed that it is difficult to estimate the reflectivity of the surface of an unfamiliar environment. In our experiments, it is possible that the observers could distinguish between unfamiliar (low and high diffuseness) and familiar (medium diffuseness) surface conditions. However, it is unclear which is more important for appearance perception: memory information or a near-daily environment. Therefore, further validation is required for the present findings.
The matte polyresin, wood, and fur stimuli were judged to be matte by impression evaluation in the pre-observation environment; impression evaluation for the wood and fur objects did not vary significantly under any diffuseness condition. Olkkonen et al. [12] indicated that the impression of matte objects remained constant under varying diffuseness conditions, and our results exhibited a similar trend. However, unlike wood and fur, the matte polyresin ball showed similar tendencies to the glossy objects. Therefore, we measured the luminance information in six directions obtained from 45° incident light for the glossy polyresin ball, matte polyresin ball, and wooden cube, using a CM-M6 multi-angle spectrophotometer (KONICA MINOLTA). The measured values are listed in Table 6. The ratio of the maximum value to the minimum value is the smallest for the wooden cube and largest for the glossy polyresin ball. Overall, matte polyresin ball scores are similar to glossy polyresin ball. Therefore, the observer was able to obtain similar information about light reflection.
Table 6. Luminance Information of the Glossy Polyresin Ball, Matte Polyresin Ball, and Wooden Cube
These results show that the appearance of glossy materials is sensitive to diffuseness. Kiyasu et al. [22] showed that the impression of glossy objects changes significantly with illumination at different diffuseness levels. Our study showed that diffuseness affects not only impression evaluation but also the fidelity and ideality of an object’s appearance. This suggests that we need to carefully consider the distribution of light, especially when dealing with the appearance of glossy objects.
6. CONCLUSION
We investigated the diffuseness of illumination suitable for reproducing the surface appearance of objects. Observers viewed objects with various appearances and evaluated the fidelity and ideality of the object appearance and impressions of the objects under four or five diffuseness conditions.
In Experiment I, we experimented with four different diffuseness conditions. We found that object appearance was faithfully reproduced under moderate diffuseness, and that diffuseness for ideal reproduction depended on the material of the object. In Experiment II, we experimented with five different diffuseness conditions, using a wider range than in Experiment I. Moderate diffuseness are rated higher for both fidelity and ideality of object appearance. In Experiment III, we used objects of the same material with different levels of surface gloss, and objects with a metallic gloss. Moderate diffuseness was again rated high in terms of both fidelity and ideality of the appearance of the objects.
The results of our study indicate that a moderate level of diffuseness reproduces the object surface appearance most faithfully and ideally. The results also suggest that the diffuseness condition that ideally reproduces the object surface appearance differs depending on the surface material properties.
Funding
Japan Society for the Promotion of Science (JP16K00368, JP19H04196).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
REFERENCES
1. W. Fujisaki, M. Tokita, and K. Kariya, “Perception of the material properties of wood based on vision, audition, and touch,” Vis. Res. 109, 185–200 (2015). [CrossRef]
2. Y.-X. Ho, M. S. Landy, and L. T. Maloney, “Conjoint measurement of gloss and surface texture,” Psychol. Sci. 19, 196–204 (2008). [CrossRef]
3. B. Xiao, W. Bi, X. Jia, H. Wei, and E. H. Adelson, “Can you see what you feel? Color and folding properties affect visual–tactile material discrimination of fabrics,” J. Vis. 16(3), 34 (2016). [CrossRef]
4. L. T. Maloney, H. E. Gerhard, H. Boyaci, and K. Doerschner, “Surface color perception and light field estimation in 3D scenes,” Vision in 3D Environments 1(3), 280–307 (2011). [CrossRef]
5. M. Ennio, T. T. James, and J. F. Norman, “Perception of lightness of glossy surfaces,” Proc. SPIE 4299, 302–311 (2001). [CrossRef]
6. R. W. Fleming, R. O. Dror, and E. H. Adelson, “Surface reflectance estimation under unknown natural illumination,” J. Vis. 1(3), 43 (2003). [CrossRef]
7. B. Hartung and D. Kersten, “Distinguishing shiny from matte,” J. Vis. 2(7), 551 (2002). [CrossRef]
8. L. T. Maloney, “Illuminant estimation as cue combination,” J. Vis. 2(6), 493–504 (2002). [CrossRef]
9. M. S. Langer and H. H. Bulthoff, “A prior for global convexity in local shape-from-shading,” Perception 30, 403–410 (2001). [CrossRef]
10. G. Obein, K. Knoblauch, and F. Vienot, “Difference scaling of gloss: Nonlinearity, binocularity, and constancy,” J. Vis. 4(9), 711–720 (2004). [CrossRef]
11. S. F. Te Pas and S. C. Pont, “A comparison of material and illumination discrimination performance for real rough, real smooth and computer generated smooth spheres,” in Proceedings of the 2nd Symposium on Applied Perception in Graphics and Visualization (2005), pp. 75–81.
12. M. Olkkonen and D. H. Brainard, “Perceived glossiness and lightness under real-world illumination,” J. Vis. 10(9):5, 1–19 (2010). [CrossRef]
13. Y. Mizokami, Y. Kiyasu, and H. Yaguchi, “Change in the appearance of objects according to the ratio of direct and diffusive light,” in CIE x046: 2019 Proceedings of the 29th CIE SESSION, Washington, DC, USA (2019), paper PP05.
14. S. C. Pont and S. F. Te Pas, “Material—illumination ambiguities and the perception of solid objects,” Perception 35, 1331–1350 (2006). [CrossRef]
15. T. Yamazoe, T. Funaki, Y. Kiyasu, and Y. Mizokami, “Evaluation of material appearance under different spotlight distributions compared to natural illumination,” J. Imaging 5, 31 (2019). [CrossRef]
16. C. Cuttle, “Research note: a practical approach to cubic illuminance measurement,” Light. Res. Technol. 46, 31–34 (2014). [CrossRef]
17. L. Xia, S. C. Pont, and I. Heynderickx, “Diffuseness metric part 1: theory,” Light. Res. Technol. 49, 411–427 (2016). [CrossRef]
18. C. Hiramatsu, N. Goda, and H. Komatsu, “Transformation from image-based to perceptual representation of materials along the human ventral visual pathway,” Neuroimage 57, 482–494 (2011). [CrossRef]
19. R. W. Fleming, C. Wiebel, and K. Gegenfurtner, “Perceptual qualities and material classes,” J. Vis. 13(8), 9 (2013). [CrossRef]
20. J. J. Koenderink, S. C. Pont, A. J. Van Doorn, A. M. L. Kappers, and J. T. Todd, “The visual light field,” Perception 36, 1595–1610 (2007). [CrossRef]
21. R. W. Fleming, R. O. Dror, and E. H. Adelson, “Real-world illumination and the perception of surface reflectance properties,” J. Vis. 3(5), 3 (2003). [CrossRef]
22. Y. Kiyasu, Y. Mizokami, and H. Yaguchi, “How direct light and diffusive light influence the appearance of objects?” J. Vis. 19(8), 54 (2019). [CrossRef]
