| Sažetak | Cilj istraživanja provedenih u ovoj disertaciji je definirati smjernice za pripremu digitalne
slike u sustavima koji generiraju sadržaje namijenjene prikazu na web stranici i za ispis na
papiru. Priprema fotografije za objavu u web okruženju podrazumijeva optimizaciju zapisa
kako bi se dobio optimalni omjer kvalitete i veličine slikovne datoteke. Ključni korak
optimizacije je redukcija broja piksela koji čine sliku, a potom slijedi sažimanje slikovnih
podataka primjenom odabranog algoritma. Očekivane smjernice dobivene su kvalitativnim i
kvantitativnim ispitivanjem relevantnih grafičkih formata i pripadajućih metrika kompresije te
određivanjem optimalnih postavki pri izradi slika za spomenuta dva različita slučaja primjene.
Razmatrana su dva grafička formata zapisa slike, JPEG i WebP. JPEG format zapisa slike
koristi se već dugi niz godina, a iako postoji niz novijih formata, i dalje se često koristi zbog
svoje široke podrške i jednostavnosti korištenja. WebP format, s druge strane, je noviji format
koji se razvijao s ciljem poboljšanja učinkovitosti kompresije slika i brzine učitavanja web
stranica. Prema dosadašnjim istraživanjima, WebP format je napredniji od JPEG-a u kontekstu
mogućnosti, no i u samoj optimizaciji slikovnog zapisa. To se pokazalo i u ovoj disertaciji.
Poseban naglasak postavljen je na razmatranje utjecaja rezolucije i kompresije na
subjektivni doživljaj kvalitete slike te utjecaj optimizacije na reprodukciju boja i veličinu
slikovne datoteke. Subjektivne metode procjene najpouzdaniji su načini mjerenja obzirom da
je ljudski vizualni sustav krajnji primatelj slikovnih informacija. Takva prosudba ocjene
temeljena na percepciji promatrača ima važnu ulogu u sustavima vizualnih komunikacija.
Kako bi se znanstvenom metodom utvrdile činjenice na temelju kojih bi se nedvosmisleno
i sveobuhvatno donijeli relevantni zaključci, razvijena je nova eksperimentalna metoda za
ispitivanje kvalitativnih i kvantitativnih karakteristika grafičkih formata na velikom broju
uzoraka, koja obuhvaća subjektivne i objektivne pokazatelje reprodukcije slike u razmatranim
okruženjima. Metodologija prikazana u eksperimentalnom dijelu rada obuhvaća niz
jednostavno ponovljivih koraka koji su potpomognuti razvijenom programskom podrškom u
vidu raznih programa čija je namjena automatizacija kompliciranih zadataka. Provedenim
istraživanja potvrđene su sve predviđene hipoteze, uspješno su ostvareni svi očekivani
znanstveni doprinosi i ostvaren je cilj doktorskog rada, odnosno definirane su smjernice za
grafičku pripremu digitalnih slika u razmatranom kontekstu. |
| Sažetak (engleski) | The aim of the research conducted in this dissertation is to define guidelines for preparing
digital images in systems that generate content intended for display on websites and for printing
on paper. Preparing a photograph for publication in a web environment involves optimizing the
file to achieve the optimal balance between quality and file size. The key optimization step
involves reducing the number of pixels that make up the image, followed by compressing the
image data using a selected algorithm. The expected guidelines were obtained through
qualitative and quantitative examination of relevant graphic formats and compression metrics,
as well as by determining the optimal settings for creating images for the two aforementioned
different application scenarios.
Optimizing images for web and print requires a special approach due to the different
characteristics of the end media. While web optimization focuses on transmission speed and
on-screen display, high resolution and detail are crucial for printing. Therefore, it is important
to adjust compression settings and image resolution according to the specificities of each
medium. Additionally, the research identified optimal compression methods and file formats
that ensure minimal loss of image quality. The combination of qualitative and quantitative
analyses has facilitated the development of concrete guidelines to streamline the process of
preparing digital images for various types of publications.
Two graphic file formats were considered: JPEG and WebP. JPEG has been used for
many years and, despite the existence of newer formats, it is still frequently used due to its
widespread support and ease of use. On the other hand, WebP is a newer format developed to
improve image compression efficiency and web page loading speed. According to previous
research, WebP format is more advanced than JPEG in terms of capabilities and image
compression optimization, a finding supported by this dissertation.
The research extensively analyzed the characteristics of both formats across various
aspects, including file size, image quality, and loading speed. The results confirmed that WebP
format provides significantly better compression with minimal loss of quality compared to
JPEG. Moreover, WebP stands out for its optimization capabilities for web environments,
making it an attractive choice for digital content intended for online display. This dissertation
provides further confirmation that WebP format is promising and could become the standard
format for web graphics in the future.
Special emphasis was placed on considering the impact of resolution and compression on
the subjective perception of image quality, as well as the impact of optimization on colorreproduction and file size. Subjective evaluation methods are the most reliable means of
measurement since the human visual system is the end recipient of visual information. Such
subjective assessment based on observer perception plays a crucial role in visual
communication systems.
The research employed various subjective evaluation methods, including participant
ratings and visual tests, to gain deeper insights into the perception of image quality in different
scenarios. The results showed that even small changes in resolution and compression can
significantly affect viewers' perception of image quality. Furthermore, optimizing resolution
and file size are crucial factors in achieving an optimal balance between visual appeal and data
transmission efficiency. Understanding the subjective aspects of image quality is essential for
developing effective optimization strategies and improving user experience in digital media.
To scientifically establish facts upon which relevant conclusions could be unambiguously
and comprehensively drawn, a new experimental method was developed to examine the
qualitative and quantitative characteristics of graphic formats on a large number of samples,
encompassing subjective and objective indicators of image reproduction in the considered
environments. The methodology presented in the experimental section of the work includes a
series of easily replicable steps aided by developed software support for automating complex
tasks. The following research hypotheses were formulated:
H1: In the context of image display on a monitor screen, image quality will be perceived
as high when the number of pixels constituting the image is greater than the number of pixels
that fit within the frame in which the image is displayed.
H2: It is possible to significantly reduce the size of the image file by combining resolution
reduction optimization with compression, while maintaining high subjective perception of
quality and accurate color reproduction.
H3: Optimizations performed in the context of image display on a monitor screen are
applicable to image preparation for printing, while maintaining a roughly equal high subjective
perception of quality between optimized and unoptimized printed images.
Additionally, anticipated and expected scientific contributions are outlined:
C1: A new experimental method for examining the qualitative and quantitative
characteristics of graphic formats on a large number of samples, encompassing subjective and
objective indicators of image reproduction in the considered environments.
C2: Guidelines for graphic preparation of digital images in systems generating image
content intended for two different environments, digital (websites) and printing.C3: Determination of optimal resolution values and compression levels in the process of
graphic image preparation using JPEG and WebP format for modern multimedia environments.
The conducted research confirmed all predicted hypotheses, successfully achieved all
expected scientific contributions, and accomplished the goal of the doctoral thesis, i.e., defining
guidelines for graphic preparation of digital images in the considered context. |
