Luminance and Color Correction for Display Stitching in Semi-Cave

Virtual Reality

Dariusz Sawicki

1 a

, Łukasz Izdebski

, Agnieszka Wolska

2 b

and Mariusz Wisełka

2 c

Warsaw University of Technology, Institute of Theory of Electrical Engineering, Measurements and Information Systems,

Warsaw, Poland

Central Institute for Labour Protection, National Research Institute (CIOP-PIB), Warsaw, Poland

Keywords: Image Stitching, Multimedia, Virtual Environment, Immersive VR, Cave.

Abstract: The most spectacular example of a virtual reality (VR) environment is Cave (Cave Automatic Virtual

Environment). Image stitching is an essential problem encountered in displaying images in the Cave VR. We

analyzed this problem in the Semi-Cave installation, a low-budget representative of the Cave VR. Seamless

stitching of displayed images requires two independent tasks: geometrical correction and color/luminance

correction. The aim of this work is to present the main aspects and the methods used for color/luminance

correction for seamless stitching of displayed images in the Semi-Cave installation. The proposed procedure

and the software developed for color correction of images were tested and verified. The final effect of

displaying stitched images was subjectively assessed. The impression of immersion into the Semi-Cave VR

was sensed by subjects, and in this way, the correctness of the proposed method was confirmed.

1 INTRODUCTION

The quality of the displayed images is a key issue in

Cave (Cave Automatic Virtual Environment) that

determines the correctness of immersion into the

created VR environment. Immersion into the VR is

understood as a specific concept that defines how

well the VR environment represents the real world

and how well it is perceived (Slater, 2003). The

following parameters are considered as the most

important in the immersion process: correctness of

color and image geometry in the stitched images,

correctness of the color rendering process, perception

of projection, and image resolution (Slater, 2003).

Image stitching is an essential problem

encountered in displaying images in the Cave VR

installation, as well as in many other multimedia

applications. We analyzed this problem in the Semi-

Cave, a low-budget representative of the Cave VR

installation. Image stitching in the Semi-Cave

installation requires two independent corrections:

geometric correction which is the first task and

described in the work (Sawicki et al., 2018); and

https://orcid.org/0000-0003-3990-0121

https://orcid.org/0000-0003-3912-605X

https://orcid.org/0000-0002-7145-6457

color/luminance correction—assuming that the

images are already geometrically corrected.

The need to correct the color of the stitched

images in the Semi-Cave installation results from the

differences in the colors of the displayed images. This

is mainly related to the differences in image

displaying by particular projectors (individual

differences and aging of the equipment). The

difference in color coordinates is a measure of the

color mismatches (Mokrzycki and Tatol, 2011).

The problem encountered in seamless stitching of

images to create a panorama is well described in the

literature (Singh and Saravanan, 2017, Pravenaa and

Menaka, 2016). Geometric correction and color

correction work together in most of the stitching

methods for creating panoramas (Pravenaa and

Menaka, 2016). Nevertheless, advanced methods of

color correction are applied in such cases (Bellavia

and Colombo, 2018). Color stitching for the Semi-

Cave VR requires not only an advanced color

correction method but also solving additional

measurement problems that we would like to

highlight here.

Sawicki, D., Izdebski, Ł., Wolska, A. and Wisełka, M.

Luminance and Color Correction for Display Stitching in Semi-Cave Virtual Reality.

DOI: 10.5220/0008168501370144

In Proceedings of the 3rd International Conference on Computer-Human Interaction Research and Applications (CHIRA 2019), pages 137-144

ISBN: 978-989-758-376-6

137

The aim of this paper is to present the main

aspects of color/luminance correction and the

procedure used for color correction for seamless

stitching of images in the Semi-Cave installation.

2 TECHNICAL ASPECTS AND

PRELIMINARY ANALYSIS

The Semi-Cave was implemented in a rectangular

room, measuring 8.6 m × 4.3 m × 6 m. The image was

created by direct projection onto the four walls of the

room. Image creation was ensured by six projectors

(Canon, XEED WUX400ST) presented in Figure 1.

In the computer system responsible for image

control, three graphic cards (NVidia GTX 980 Strix

OC SLI) were used. Each of the graphic cards

supported two projectors. The whole system ensured

a smooth and sharp display of images at a resolution

of 1920 × 1080. The details of the installation are

described in the work (Sawicki et al. 2017).

Figure 1: Arrangement of projectors in the Semi-Cave

laboratory (Sawicki et al., 2018).

The way in which the images are displayed in

Semi-Cave strongly influences the possibilities of

geometric and color corrections. Stitching of images

should be done differently on the same wall (e.g.,

Image_1 and Image_2 in Figure 1) and on

neighboring walls, respectively (e.g., Image_2 and

Image_3 in Figure 1). In the first case, for the correct

display and ease of correction, it is convenient to use

overlapping images. For geometric correction, an

overlapping part with a width of 50 pixels is

considered. In the second case (in the corner of the

room), the images are corrected at the place of

“contact” (i.e., at the edge of the wall corner). If one

subsystem of display correction works for geometry

as well as for color, the adopted rules must be applied

in both the cases. As a consequence, the color needs

to be equalized in the overlapping part. This problem

does not occur in the case of geometry correction—

the reproduction of properly corrected fragments in

the same place does not affect the geometry, but

changes the luminance or color.

On the basis of the analysis of the entire correction

process (geometric and color correction), we assumed

that the geometry correction should be performed as

the first task. This is due to the fact that it does not

require additional measurements but only the visual

assessment of corrected and displayed images.

However, color/luminance correction requires taking

measurements during correction.

3 COLOR SPACE USED IN

SEMI-CAVE

The projectors used enable working in many different

modes (Canon User’s Manual, 2013). However, the

manufacturer declares compliance with the sRGB

space only for the Photo/sRGB mode.

The sRGB color space is defined in the IEC

standard (IEC 61966-2-1:1999, 1999). Linearization

(gamma) dependencies for each coordinate (R, G, B)

are described by equations (1)–(4). Due to the

additivity of the color components of radiation, the

corresponding dependence will also apply to the

black–white scale (BW—the scale of gray levels).

for R

≤ 0.0031308



  



  



  



(1)

for R

> 0.0031308



 







 

  







 

 







 

(2)

for R,G,B ≤ 0.04045





















(3)

for R,G,B > 0.04045







 







 







 



(4)

Where R,G,B  [0,1] and R

 [0,1]. R,G,B are

the color coordinates that define a color in the sRGB

space, for example, for gray scale from black

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

138

(R=G=B=0) to white (R=G=B=1), while R

are

the luminance coordinates of the displayed image.

The dependence between the image luminance

and control parameters (stimulation) in monitors is

described by the power function. At the same time,

the dependence of the brightness of the perceived

image is also not linear and can be described by a

similar function—the perceptual dependencies of the

sense of sight of humans are described by the Weber–

Fechner’s law (Bruce, 2014). Hence, the

mathematical description of nonlinear dependencies

(1)–(4) is an attempt to match the technological

properties of the equipment to the perceptual abilities

of the human. The assumption of an appropriate

function describing the nonlinear relations very

strongly affects the perception of differences. In the

software used for controlling graphic devices as well

as for processing images and computer graphics, the

coordinates R,G,B are most often expressed in a

binary form (2

bit_depth

-1). For example, for

bit_depth = 8 the coordinates change in the range of

[0,255]. This requires rescaling from the binary form

to the form expressed in formulas (1)–(4) or vice

versa.

It is worth noting that according to the IEC

standard (IEC 61966-2-1:1999, 1999), the power in

equations (2) and (4) has a value of 2.4 instead of 2.2,

that is, the value assumed to be typical in the gamma

correction of PC hardware (Poynton, 2005, Poynton,

2012). In practice, the approximate formulas (5) and

(6) are most often used.

for R,G,B [0, 1] and R

[0, 1]



























(5)













































(6)

4 MEASUREMENTS AND

PERCEPTUAL

IDENTIFICATION OF COLOR

DIFFERENCES IN SEMI-CAVE

A Chroma Meter CS-200 instrument (Konica Minolta

(Chroma Meter CS-200, 2013) was used to measure

the luminance and color of the displayed images. The

instrument is adjusted to V()—the relative spectral

luminous efficacy of the eye adapted to brightness.

This allows the direct measurements of luminance

and X,Y,Z color coordinates in the CIE XYZ color

space. However, it does not allow color correction for

usage in the display software. In addition, the

luminance values of the displayed images in Semi-

Cave are at a low range (below 60 cd/m

), which

further reduces the usefulness of this type of

measurement in color correction.

A number of experiments were carried out which

confirmed that the X,Y,Z color coordinates cannot be

determined directly in an efficient and useful way.

Ultimately, a decision was made to carry out the

measurements and color correction indirectly and

independently for each R, G, and B component. In this

case, the measured luminance value will indeed take

into account the curve V(). This is enough to

compare the image fields displayed by the

neighboring projectors, but within only one

component. The display correction will be based on

the determined difference. This will allow displaying

identical colors regardless of how the projectors

display them. On the other hand, this means treating

the entire displaying process (software, graphic cards,

projectors, wall reflectance) as a “black box.” We

know the input parameters (binary values describing

the color of an image), which are defined in the

software and are subject to color correction, and the

output parameters (information displayed on the

Semi-Cave wall), which are subject to measurement.

The aim of the correction software is to adjust the

input parameters according to the changes needed in

the output parameters and take the relevant

measurements.

5 COLOR CALIBRATION IN

SEMI-CAVE: ANALYSIS OF

POSSIBILITIES

For system calibration, a set of color images was

prepared in four groups: for each component R, G, B,

and BW (gray levels). A comparison of all the images

and their corresponding R,G,B color coordinates is

summarized in Table 1. However, unfortunately, the

changes in the color of low-luminance images cannot

be actually recognized in a printout of this table—

especially in the case of binary components R (31,0,0)

and B (0,0,31)—due to limited printing capabilities.

The analysis of the results of the preliminary tests

leads to quite interesting conclusions. On the one

hand, the results are in line with expectations—with

Luminance and Color Correction for Display Stitching in Semi-Cave Virtual Reality

139

known perceptual properties of the human sight. A

human cannot identify differences at a level of 1 bit

(1/256 of the displayed full scale of color) for any

component in the dark parts of an image. It is assumed

that a human can distinguish a maximum of 60–90

shades of one color. In the bright parts of a picture, a

human can distinguish the shades of red color better

than the shades of blue, and in an exceptional

situation of a specific color neighborhood, a

difference at the level of 1 bit would be noticed. On

the other hand, the measurements of luminance in

many of the analyzed cases do not help in

distinguishing the images. This is due to the reflection

properties of the walls in the Semi-Cave. The use of

reflective paint for a projector wall does not enable as

high reflectance as a professional projection screen.

Table 1: The set of used images (R,G,B and BW—gray)

with proper R,G,B binary coordinates.

R image

binary

coordinates

G image

binary

coordinates

B image

binary

coordinates

BW image

binary

coordinates

0,0,0

31,0,0

0,31,0

0,0,31

31,31,31

63,0,0

0,63,0

0,0,63

63,63,63

127,0,0

0,127,0

0,0,127

127,127,127

191,0,0

0,191,0

0,0,191

191,191,191

223,0,0

0,223,0

0,0,223

223,223,223

255,0,0

0,255,0

0,0,255

255,255,255

Color interpolation at a resolution of 1 bit for each

component corresponds to the border perceptual and

measurement possibilities of identifying color

differences. The perceptual capabilities of a human

do not allow identifying differences at such a

resolution in the whole range of luminance that can

be obtained while displaying images. Nevertheless,

experiments have shown that under real conditions in

the Semi-Cave, differences at a level of 1 bit may be

noticed in specific displayed images.

Color correction means interpolation of the values

of each color coordinate in a specific area of a

displayed image. If the side edge of a displayed

rectangle is the place of matching the common color,

it can be assumed that the area on which the

interpolation will be carried out will be a rectangle.

Its height will be consistent with the height of the

displayed image, with one edge being the edge of the

match and the other edge determining the “depth of

penetration” of interpolation in the area of the

displayed image.

Two interpolation parameters need to be defined.

The first is the “depth of penetration.” After

conducting preliminary experiments, it was assumed

that the interpolation would cover one-third area of

the surface of the displayed image. This means that

each image should be divided into three equal parts

(left, central, and right). The luminance and color will

not be modified in the central part, but will be

modified in the area on the left side or right side

where interpolation is associated with the left edge or

right edge of the match, respectively.

The second parameter is the “type of changes”

carried out in the value of the interpolated parameter.

The simplest solution is to change linearly the values

of color coordinates in the stitched images. The

problem seems to be trivial: the color correction of

stitched images is done in a large space and the

differences in corrected colors are small. In such a

large space, local, high color changes may occur—

depending on the content of the image. Therefore, the

differences related to color correction would not be

noticeable. However, in many cases, we can see large

surfaces with almost the same color—for example,

the sky with a similar or the same blue color. In this

situation, because of human perception, incorrect

local color changes will be unnatural.

The simplest linear change is practically

unfavorable in every case. Taking into account the

results of studies on perception and measurement in

real conditions, we assumed that the curve of color

changes resulting from the correction should be

smooth (in the sense of continuity of the first

derivative). This is important because of lateral

inhibition (Bakshi et al., 2017, Hall, 1989), a

phenomenon that causes even the smallest local

unevenness of changes in color or luminance to be

emphasized and perceived by the sense of sight. It is

worth considering the appropriate connecting curve

used in computer graphics and animation—curves

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

140

derived from the set of easing functions (Penner,

2002, Izdebski and Sawicki, 2016). This means that

to describe the changes in the values of color

coordinates, the appropriate (smooth) function should

be used—for a “smooth start and smooth stop.” The

simplest functions that produce good results are the

polynomial functions of InOutQuad and InOutCubic

or non-polynomial InOutSin from the set of easing

functions (Penner, 2002, Izdebski and Sawicki, 2016)

— Figure 2.

Figure 2: The InOutCubic easing function — an example of

function with a “smooth start and smooth stop” (Izdebski

and Sawicki, 2016).

6 THE RESULTS: CORRECTION

PROCEDURE

Analysis of the results obtained from the conducted

experiments allows proposing a simple procedure for

color correction for stitching images in Semi-Cave.

Correction is carried out independently for R, G, and

B, assuming that in the case of each component, the

other two components are zero. This procedure is

based on a comparison of luminance in specific fields

of adjacent (stitched) images and is carried out in the

following three steps:

S1. Measure the luminance of neighboring (adjacent)

fields (Figure 3).

S2. If the measured luminances have the same value,

finish the correction and stop the procedure; if not,

perform operation S3.

S3. In the correction program, change the displayed

color—by inserting the value of an appropriate

correction factor in the color coordinate. Most

often, it is enough to change the value of the factor

in only one of the neighboring fields, but if the

differences are large, change the correction

factors in both the fields, and return to S1.

The measurements carried out confirmed the

possibility of using approximate formulas in versions

(5) and (6) in real conditions of Semi-Cave.

Therefore, it is possible to propose a simple method

for determining the value of correction factors in the

proposed procedure of color correction.

Figure 3: The image generated by the correction program,

for example, for color G (0,255,0). Correction will take

place along the left or right edge—where identical images

will be displayed. The colors of the neighboring images are

compared along the joining edge in three fields: upper,

lower, and middle. The middle field corresponds to the

height of the “horizon” of the displayed image. It was

assumed that at such a height, the displayed information is

the most important.

Let L1

be the value of the measured luminance in

the first field (e.g., the image on the left side of the

joining edge), where k is one of the components (R,

G, or B) or corresponds to the correction of gray

levels (BW—in this case, all three components are

corrected equally in the program). For example, L1

is the measured luminance of the red component and

corresponds to R

in equations (1)–(6). Similarly, let

be the value of the measured luminance in the

second field (e.g., the image on the right side of the

joining edge). Let us assume that the correction factor

W for the first field will be introduced in the color

correction program. This means that L2

should be the

expected luminance value after correction in the first

field. Taking into account equation (5), the value of

W can be determined on the basis of equation (7).

































(7)

For small differences in luminance values (a

maximum difference of 20%–30%—which

practically always occurs in Semi-Cave), equation (7)

can be represented in a simple approximated form (8)

using the power series expansion.

   





 







(8)

Luminance and Color Correction for Display Stitching in Semi-Cave Virtual Reality

141

The simple correction procedure with feedback

proposed above requires consideration of specific

cases. The full procedure is as follows:

P1. Check if the display geometry is correct. If not,

perform the geometry correction. Select the

two stitched images for color correction (called

in the procedure as the left image and the right

image).

P2. Carry out the correction independently for the

components R, G and B. Repeat the procedure

independently for the appropriate components.

The order is not important.

P2.1. Select a component (R or G or B).

P2.2. Specify the level of color (binary) displayed on

the control screen. As a standard, the color at the level

of R (127,0,0) or G (0,127,0) or B (0,0,127) will be

displayed. In this way, the luminance of the displayed

information corresponds to approximately 18% of the

full range of luminance, which is good enough for

assessing the lighting in typical scenes

P2.3. Display the same control image with the

specified color and specified level on both stitched

images.

P2.4. Measure the luminance in the adjacent fields on

both sides of the joining edge (on the left and right

stitched image independently). Compare the results of

the measurement.

P2.5. Apply display correction (correction factor)

according to formula (8). If the difference in

luminance of measurements between images (left and

right) is small, apply correction only for one image

(either left or right). If the difference is greater (above

15%), apply the correction for both images.

P2.6. Repeat the luminance measurement. If the

difference is aligned, finish the correction for a given

pair of fields (or for a given component), if there are

still measurable differences, repeat operations P2.4 –

P2.5.

P3. Display both the images in a standard view (such

that the entire surface of the image is visible, not just

the separate squares). Rate visually the stitched

images. Evaluate the adjacent fields at different

heights of the joining edges. If still there are

perceptual or measurement differences, repeat the

whole P2 operation of correction procedure.

On the basis of the procedure proposed for color

correction, an application carrying out appropriate

tasks has been prepared. The software has been

implemented in Visual Studio using the Vulkan

libraries (Sellers and Kessenich, 2016) and OpenGL

(Sellers et al., 2015), and works on the level of shader

drivers (Bailey and Cunningham, 2011). The

application works in two modes: edit mode and work

mode. Both modes use a common code associated

with managing the display of windows in a Windows

environment.

Operations of color correction were combined

with previously developed program for geometric

correction (Sawicki et al., 2018), thereby creating one

subsystem for display correction.

7 VERIFICATION OF THE

SOLUTION

The method developed for color correction was tested

using selected color images. The tests verified the

color composition between the images, which had

different degrees of complexity and showed color

variations. The comparison was based both on

perceptual verification and on the luminance

measurements obtained for each component.

Figure 4: An example of stitching two images with a large

color mismatch between components. The images were

combined with up-down and down-up changes between R

and G: from (255,127,127) to (127,255,127). a) Stitched

images before correction. b) After correction.

Preliminary stitching tests carried out with the

images with color correction within each color

component showed positive results. The stitching

tests were further carried out with specially prepared

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

142

images that had very large color mismatches both

within one component and between two components.

An example of image correction is presented in

Figure 4. The stitching tests were carried out to

confirm the correctness of the operation of the

developed correction system. Changes in the

correction level (or changes in the values of

correction factors) are subject to linear dependencies.

The correction problem is scalable – can be scaled to

the range allowed by the maximum hardware settings.

Therefore, if it is possible to achieve adequate

correction for selected (very large) color differences,

it will also be possible to achieve proper correction

for all smaller color differences

Figure 5: An example view of the presented VR

environment generated in Semi-Cave. The photo was taken

from a point consistent with the position of the observer in

VR.

To check the operation of the entire subsystem

developed for the correction of both geometry and

color in Semi-Cave, we used UNITY to create a VR

environment. The principles of perspective projection

for acquiring images were determined, taking into

account the size and shape of the laboratory. It was

assumed that the observer (projection center) is

located exactly in the center of the room. Images

created using the methods of computer graphics allow

defining elements with an accuracy of 1 pixel. In

addition, computer graphics facilitate changing the

projection conditions and display conditions. This

approach allows visual evaluation of the correctness

of the display. In our case, visual evaluation is

considered the best because it gives an opportunity to

evaluate the perception of images and immersion into

a VR environment. We carried out the evaluation with

10 participants (other than authors). The conducted

tests confirmed the correctness of the operation of the

display system. A sample picture taken in Semi-Cave,

using Sponza Palace in Dubrovnik as a model to

create a VR, is shown in Figure 5.

8 SUMMARY

Image stitching in the Semi-Cave installation is a

complex issue involving several independent

problems. The subsystem developed for the

correction of geometry and color is an individual

solution adapted to the conditions of Semi-Cave. In

addition, our software used for display driver allows

for full control of the entire process of displaying

information.

Color correction is a much more difficult task than

geometry correction. It requires obtaining a series of

measurements under strictly defined conditions. In

addition, it requires the use of a specially prepared set

of control images. Therefore, the solution developed

both indicates the specific methods for conducting

correction and helps prepare the environment to carry

out this correction.

The analysis of the process of displaying

information in real conditions of Semi-Cave showed

that the best solution would be to perform an indirect

comparison of colors displayed in the neighboring

images. For this, control images with colors

corresponding to a particular R, G, B component were

used, and the luminance measurement was carried

out. The measurement obtained helped in carrying out

the color correction in a convenient way.

The tests carried out showed that the developed

method allows achieving seamless stitching of

images in terms of their color matching. The method

can give good results even if an extreme color

mismatch existed before correction. The entire

subsystem (for the correction of geometry and color)

was also tested using the VR environment created for

Semi-Cave. It was assumed that the observer’s

location and perspective projection (in the VR) were

consistent with the shape and size of the real room.

The best test in this situation is the perceptual

assessment and the impression of immersion into the

VR. The experiments carried out confirmed the

correctness of the operation of the developed system.

It is worth comparing our method to solutions

known from the literature. The installation conditions

of Semi-Cave allowed performing the geometry

correction independently as the first step. In this way,

we could assume that the images are already

geometrically corrected—this means that it has

simplified the task in relation to advanced

contemporary methods (Bellavia and Colombo,

2018). On the other hand, the task turned out to be

much more difficult because of the need to measure

the luminance/color (and its components)—a

problem that does not occur with the typical stitching

of the panorama (Bellavia and Colombo, 2018).

Luminance and Color Correction for Display Stitching in Semi-Cave Virtual Reality

143

The original method of seamless stitching of

images in the Semi-Cave installation is presented in

this paper. However, we only consider it as a

technical tool necessary to create a convincing VR

environment for conducting scientific research in the

future. We are planning to create a VR environment

which will help reproduce images of workplaces,

garden, forest, etc. This will allow examining the

impact of the visual environment (objects, colors,

contrasts, etc.) on the cognitive performance and the

well-being of the subjects.

ACKNOWLEDGEMENTS

This paper has been based on the results of a research

task carried out within the scope of the fourth stage of

the National Programme "Improvement of safety and

working conditions" partly supported in 2017–2019

within the scope of research and development --- by

the Ministry of Science and Higher Education /

National Centre for Research and Development. The

Central Institute for Labour Protection -- National

Research Institute is the Programme's main

coordinator.

REFERENCES

Bailey, M., Cunningham, S., 2011. Graphics Shaders:

Theory and Practice, Second Edition. A K Peters/CRC

Press.

Bakshi, A., Ghosh, K., 2017. A Neural Model of Attention

and Feedback for Computing Perceived Brightness in

Vision. In: Handbook of Neural Computation. Chapter

26, pp.487-513. Academic press. https://doi.org/

10.1016/B978-0-12-811318-9.00026-0.

Bellavia, F., and Colombo, C., 2018. Dissecting and

Reassembling Color Correction Algorithms for Image

Stitching, IEEE Transactions on Image Processing, 27

(2), 735-748, doi: 10.1109/TIP.2017.2757262.

Bruce, V., Georgeson M.A., Green, P.R., 2014. Visual

Perception. Psychology Press.

Canon User’s Manual, 2013. Multimedia Projector

WUX450, WX520, WUX400ST, WX450ST. Canon

Inc. Polish Edition. Canon Europa N.V.

Chroma Meter CS-200 Konica Minolta Instruction Manual,

KONICA MINOLTA, INC. 2013. https://www.

konicaminolta.com/instruments/download/instruction_

manual/display/pdf/cs-200_instruction_eng.pdf. last

accessed 2019/04/02.

Hall, R., 1989. Illumination and Color in Computer

Generated Imagery. Springer.

IEC 61966-2-1:1999. Multimedia systems and equipment -

Colour measurement and management - Part 2-1:

Colour management - Default RGB colour space -

sRGB (https://webstore.iec.ch/publication/6169.) IEC

Webstore. International Electrotechnical Commission.

Retrieved 3 March 2017.

Izdebski, Ł., Sawicki, D., 2016. Easing functions in the new

form based on Bézier Curves, In Proc. of the

International Conference on Computer Vision and

Graphics (ICCVG 2016), Warszawa. September 19-21,

2016. LNCS 9972, 37-48. Springer. doi:10.1007/978-3-

319-46418-3_4

Mokrzycki, W.S., Tatol, M., 2011. Color difference ΔE : a

survey. Machine Graphics and Vision. 20 (4), 383-411.

Penner, R., 2002. Motion, Tweening, and Easing.

Chapter 7. In: Programming Macromedia Flash MX,

pp.191-240. McGraw-Hill/OsborneMedia. http://

robertpenner.com/easing/penner_chapter7_tweening.p

df, last accessed 6 April 2019.

Poynton Ch.A. Color technology. 03.04.2005.

http://poynton.ca/Poynton-color.html , last accessed

2018/10/02.

Poynton Ch.A., 2012. Digital Video and HD Algorithms

and Interfaces. (2nd ed.) Morgan Kaufmann.

Pravenaa, S., & Menaka, R. (2016). A methodical review

on image stitching and video stitching techniques.

International Journal of Applied Engineering

Research, 11(5), 3442-3448.

Sawicki, D., Wolska, A., Wisełka, M., Żukowski, J., Sołtan,

M., Związek, W., 2017. Semi-Cave as an example of

multimedia dedicated to study the impact of audiovisual

environment on human psychophysiology. In Proc. of

the International Conference on Computer-Human

Interaction Research and Applications (CHIRA 2017).

Funchal, Madeira, Portugal. October 31 – November 2,

2017, pp. 103-110. doi: 10.5220/0006497601030110.

Sawicki, D., Izdebski, Ł., Wolska, A., Wisełka, M., 2018.

Geometrical Picture Integration in Semi-Cave Virtual

Reality. In Proc. of the International Conference on

Computer-Human Interaction Research and

Applications (CHIRA 2018), Seville, Spain, 19-21

September, 2018, pp. 100-107. doi:10.5220/

0006922701000107.

Sellers, G., Kessenich, J., 2016. Vulkan Programming

Guide: The Official Guide to Learning Vulkan.

Addison-Wesley.

Sellers, G., Wright R.S. Jr., Haemel, N., 2015. OpenGL

Superbible: Comprehensive Tutorial and Reference

(7th Edition). Addison-Wesley Professional.

Singh, S.K., Saravanan, C., 2017. Analysis of moving DLT,

image and seam selections algorithms with MS ICE,

autostitch, and OpenCV stitcher for image stitching

applications. In Proc. of the International Conference

on Electrical, Electronics, Communication, Computer,

and Optimization Techniques (ICEECCOT), Mysuru,

2017, pp. 399-404. doi: 10.1109/ICEECCOT.

2017.8284536.

Slater, M., 2003. A Note on Presence Terminology,

http://www.cs.ucl.ac.uk/research/vr/Projects/Presencia

/ConsortiumPublications/ucl_cs_papers/presence-

terminology.htmmelslaterJan27200391557.htm , last

accessed 2017/11/09.

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

144