Rosin P., Collomosse J. (eds.) Image and Video-Based Artistic Stylisation

Springer, 2013. — 396 p.Almost two hundred years ago, the advent of photography was heralded as the beginning of the end for traditional painting. Rather than rendering painting obsolete, the technology instead motivated a new era of abstraction in visual art, delivering— among many other movements—Impressionism, Futurism, and Cubism, which continue to inspire contemporary art. Similarly, the astonishing achievements in visual realism delivered by Computer Graphics have motivated new research into the rendering of non-photorealistic styles. Non-photorealistic Rendering (NPR) is now a firmly established field within Computer Graphics, spanning over two decades of research. With origins in artistic simulation and scientific visualization, NPR has now broadened to intersect computational photography, perceptual modelling and interaction design. NPR research regularly appears in top tier graphics conferences and journals, and has delivered commercial impact through digital photography and mobile applications, and through the creative industries.
This book assembles a catalogue of classical and contemporary techniques capable of transforming 2D footage—i.e. images and video—into synthetic artwork. This sub-discipline within NPR is often referred to in the literature as Artistic Rendering, and sometimes by the more specific title Artistic Stylization. Even limiting ourselves to the rendering of images and video primarily for aesthetic value, there has been a huge diversification and development of the field over the past decade— approximately the time since the last survey of the field was published.
One significant development has been the emergence of NPR as a truly multidisciplinary field; a focal point for the convergence of Computer Graphics, Computer Vision, Human Computer Interaction and perceptual Psychology. The convergence with Computer Vision is particularly relevant to this book’s topic of 2D artistic stylization. The increasing complexity and diversity in style demanded by techniques demands a correspondingly greater degree of sophistication in the parsing and extraction of information from source footage. In the mid-1990s when automated artistic stylization techniques began to emerge, there was a reliance upon low-level image processing operators to guide the rendering process. In the early 2000s mid-level interpretation of imagery through image segmentation, perceptual salience measures, and more sophisticated filtering operators yielded improved style diversity and the robustness and temporal stability necessary to coherently stylize video. As the field matures it is now common to see a fusion of even more sophisticated image parsing, combined with careful interface design, recognizing the role of artistic stylization as a practical creative tool. Consequently in recent years, this research has begun to deliver commercial impact in major digital image and video manipulation products.Strokes, Marks and Filters for Artistic Stylization.
Stroke Based Painterly Rendering.
A Brush Stroke Synthesis Toolbox.
Halftoning and Stippling.
Non-photorealistic Shading and Hatching.
Artistic Stylization by Nonlinear Filtering.
NPR for Traditional Artistic Genres.
Stylization from Structure.
Region-Based Abstraction.
Gradient Art: Creation and Vectorization.
Depiction Using Geometric Constraints.
Artificial Mosaic Generation.
Non-photorealistic Rendering with Reduced Colour Palettes.
Artistic Rendering of Portraits.
Stylized Animations.
Temporally Coherent Video Stylization.
Computer-Assisted Repurposing of Existing Animations.
Part IV Evaluation and Impact of Artistic Stylization.
Evaluating and Validating Non-photorealistic and Illustrative Rendering.
Don’t Measure—Appreciate! NPR Seen Through the Prism of Art History.
NPR in the Wild.
Erratum to: Artistic Rendering of Portraits.