Perceptual uniformity in digital image representation and display

Poynton, Charles; Funt, Brian V.

doi:10.1002/col.21768

Cited by 22 publications

(14 citation statements)

References 25 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These three coefficients represent the intensity (luminance) perception of a standard observer trichromat human to light of the precise Rec. 709 [42] additive primary colors that are used in the definition of sRGB.…”

Section: Colorimetric Grayscale Conversionmentioning

confidence: 99%

Saliency Preservation in Low-Resolution Grayscale Images

Yohanandan

Song

Dyer

et al. 2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Visual salience detection originated over 500 million years ago and is one of nature's most efficient mechanisms. In contrast, many state-of-the-art computational saliency models are complex and inefficient. Most saliency models process high-resolution color (HC) images; however, insights into the evolutionary origins of visual salience detection suggest that achromatic low-resolution vision is essential to its speed and efficiency. Previous studies showed that lowresolution color and high-resolution grayscale images preserve saliency information. However, to our knowledge, no one has investigated whether saliency is preserved in low-resolution grayscale (LG) images. In this study, we explain the biological and computational motivation for LG, and show, through a range of human eye-tracking and computational modeling experiments, that saliency information is preserved in LG images. Moreover, we show that using LG images leads to significant speedups in model training and detection times and conclude by proposing LG images for fast and efficient salience detection.Recently, deep neural networks have achieved state-of-the-art performance on various saliency benchmarks [9,10,11,12]. Nevertheless, this success comes at high computational costs [13,14]. Training and running these networks is time-and resourceintensive, which is not easily scalable to resource-limited devices [13]. Processing highresolution or stacked multi-resolution color images is resource-intensive and contributes to this limitation [15]. In contrast, natural visual salience detection proves to be much more efficient. A deeper understanding of the evolutionary origins of visual salience detection suggests that bottom-up saliency is computed from achromatic low-resolution information [16].Previous studies have shown that low-resolution color (LC) [17,18,19] and highresolution grayscale (HG) [20,21,22,23,24,25] images preserve saliency information, yet are significantly more computationally attractive than high-resolution color (HC) images. Low-resolution grayscale (LG) images are even more computationally attractive, compared to LC and HG images. Nevertheless, to our knowledge, no one has investigated whether saliency information is preserved in LG images. In this study, we therefore investigate saliency preservation in LG images, and present the following three contributions: (1) linking low-resolution grayscale information with the bio-inspired evolutionary origins of visual saliency, (2) assessing the preservation of saliency information in low-resolution grayscale images, and (3) proposing low-resolution grayscale images for fast and efficient saliency detection. Therefore, based on a deeper understanding of the evolutionary origins of visual saliency, together with knowledge gained from studies investigating salience preservation in LC and HG images, we hypothesize that saliency information is well-preserved in LG images.

show abstract

Section: Colorimetric Grayscale Conversionmentioning

confidence: 99%

Saliency Preservation in Low-Resolution Grayscale Images

Yohanandan

Song

Dyer

et al. 2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…This means that a certain light power produces different luminous sensations depending on its wavelength. To take this into account, CIE defines the concept of luminance, which is proportional to optical power across the visible wavelengths, weighted according to a standardized spectral weighting that approximates the spectral sensitivity of normal human vision and expresses it of the following form:

Y = k_{m} {false\int}_{780 n m}^{380 n m} E (|, λ) V (|, λ) d λ

where;…”

Section: The Sensory Action Of Lightmentioning

confidence: 99%

“…Nevertheless, we will use the relative luminance with values from “0” to “100.” Where “0” corresponds with black and “100” corresponds with a reference white. The reference white value for relative luminance is chosen to correspond roughly to the luminance of a near‐perfect diffuse reflector in the scene …”

Section: The Sensory Action Of Lightmentioning

confidence: 99%

“…The reference white value for relative luminance is chosen to correspond roughly to the luminance of a near-perfect diffuse reflector in the scene. 10 However, luminance is a linear-light measure, directly proportional to light power, but our perceptive response of the luminance is not linear. Humans are much more sensitive to small light changes under low light conditions than under bright light conditions.…”

Section: The Sensory Action Of Lightmentioning

confidence: 99%

See 1 more Smart Citation

Study of balance of images using visual weight

Parada-Castellano¹

2015

Color Research & Application

View full text Add to dashboard Cite

The purpose of this research is to analyze the factors that create a sensation of balance or unbalance when a color image is observed depending on the light transmitted, reflected, and/or emitted by the objects that appear in it. This study is carried out by calculating the center of gravity (CG) of each image. The weight and the CG are obtained by calculating the force resulting from the different visual weights that make up the visual image and its position on this. When the visual CG and the geometric center coincide, it is recognized that the image is balanced. The action that acts on the balance of the lightness of an image is labeled as the visual weight. The visual weight of a figure is defined from the amount of visible light coming from it and the space occupied by this in the visual image. The colors show us the amount of light coming from a figure, while the shape provides us with information about the space that it occupies. The study explores how the qualities of appearance of threedimensional objects affect visual weight and how it can vary due to changes in light, motion of objects or the perspective of the image. Since these calculations can be very complex, we have developed the software called "PesoVisual" that automatically performs these operations on digital color images. Similarly, this software enables the analysis of the variability of the visual image.

show abstract

“…The digital files do not display uniformly across varied consumer level electronic displays using sRGB colour space, so there are practical limitations in producing a digital version. 19 Each of the 10 plates displays a ring shaped tritan foreground on a neutral background. Each ring has a gap, so the target resembles the letter C with the gap oriented to one of the eight positions marked A to H. Spectrophotometric analysis showed that the dominant wavelengths of the foreground pigments cluster around 440 nanometres.…”

Section: Introductionmentioning

confidence: 99%

The “C test” for tritan discrimination

Franzco¹,

Doba²

2017

Color Research & Application

View full text Add to dashboard Cite

The aim of this study was to develop and evaluate a clinical standard capable of generating ordinal measures of monocular tritan discrimination. A novel pseudoisochromatic plate test was developed, called the "C test." It contains 10 progressively desaturated plates arranged in perceptually distinct steps that provide a standard for analyzing the threshold for tritan discrimination. The most difficult plate that can be detected is the "C score." Relevant diagnostic findings including the C score, Logmar acuity, Pelli-Robson score, and LOCS III lens grading were prospectively recorded in 568 eyes. A total of 355 normal eyes were selected for statistical analysis. The correlations between C score, Logmar acuity, and Pelli-Robson score with LOCS III grading were analyzed with Spearman rank analysis. The strongest correlations were between the Logmar acuity with posterior subcapsular cataract (PSC) (0.68), the C score with nuclear colour (NC) (-0.58), and the Logmar acuity with NC (0.57). The lower normal limit for the C score was 9 in pseudophakes and phakic eyes with NC < 3, 7 for eyes with NC 5 3, but undetermined for eyes with NC > 3. A comparison of C score distribution between three distinct age ranges of pseudophakic eyes (n 5 136) showed no significant variation (P 5 .486). The Spearman rank correlation between C score and Logmar acuity was 0.028, and between C score and Pelli-Robson score was 0.012. The normal limit for the C score used under 800 Lux is 9 when NC < 3 and 7 when NC 5 3 in all age groups. K E Y W O R D Sacquired dyschromatopsias, C test, nuclear color, pseudoisochromatic plate test, tritan discrimination | I NT ROD UCTI ONClinical vision testing consists mainly of central achromatic resolution, peripheral luminance contrast, and red/green chromatic discrimination. These tests explore the parvocellular and magnocellular pathways but leave the koniocellular pathway largely undetermined. The Lanthony tritan album was an attempt to fill an apparent clinical need for a rapidly administered inexpensive plate test of tritan threshold, but it has serious limitations. These include a high false positive rate due to having only four choices per trial, a small range of only five plates, and inconsistent colorimetric properties between editions. 9Most colour-sorting tests use "foveal" stimuli because they are designed around the 28 field of the 1931 CIE standard observer; however, the central 0.358 of the human fovea contains no short wavelength sensitive (S) cones, and the 58 |

show abstract

Perceptual uniformity in digital image representation and display

Cited by 22 publications

References 25 publications

Saliency Preservation in Low-Resolution Grayscale Images

Saliency Preservation in Low-Resolution Grayscale Images

Study of balance of images using visual weight

The “C test” for tritan discrimination

Contact Info

Product

Resources

About