Learning style translation for the lines of a drawing

Freeman, William T.; Tenenbaum, Joshua B.; Pasztor, Egon

doi:10.1145/588272.588277

Cited by 48 publications

(22 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…10, our method performs better than the method in [5]. 8 With pose variations, the major problem of the results by [5] is to lose some structures especially on the profile. This problem can be efficiently alleviated by the shape priors, photo-tosketch patch matching and gradient compatibility designed in our model.…”

Section: Lighting and Pose Variationsmentioning

confidence: 82%

See 1 more Smart Citation

Lighting and Pose Robust Face Sketch Synthesis

Zhang

Wang

Tang

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

Abstract. Automatic face sketch synthesis has important applications in law enforcement and digital entertainment. Although great progress has been made in recent years, previous methods only work under well controlled conditions and often fail when there are variations of lighting and pose. In this paper, we propose a robust algorithm for synthesizing a face sketch from a face photo taken under a different lighting condition and in a different pose than the training set. It synthesizes local sketch patches using a multiscale Markov Random Field (MRF) model. The robustness to lighting and pose variations is achieved in three steps. Firstly, shape priors specific to facial components are introduced to reduce artifacts and distortions caused by variations of lighting and pose. Secondly, new patch descriptors and metrics which are more robust to lighting variations are used to find candidates of sketch patches given a photo patch. Lastly, a smoothing term measuring both intensity compatibility and gradient compatibility is used to match neighboring sketch patches on the MRF network more effectively. The proposed approach significantly improves the performance of the state-of-the-art method. Its effectiveness is shown through experiments on the CUHK face sketch database and celebrity photos collected from the web.

show abstract

Section: Lighting and Pose Variationsmentioning

confidence: 82%

“…Computer-based face sketch synthesis is different from line drawing generation [7] [8]. Line drawings without texture are less expressive than sketches with both contours and shading textures.…”

Section: Introductionmentioning

confidence: 99%

Lighting and Pose Robust Face Sketch Synthesis

Zhang

Wang

Tang

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

show abstract

“…Hertzmann et al [HOCS02], as well as Freeman et al [FTP03] address a similar problem. Neither method reproduces the interrelation of strokes within a pattern.…”

Section: Stroke Synthesismentioning

confidence: 96%

Stroke Pattern Analysis and Synthesis

Barla

Breslav

Thollot

et al. 2006

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…Previous methods [Hsu and Lee, 1994, Schlechtweg et al, 1998, Strothotte and Schlechtweg, 2002 used such style parameters, applied a vector or texture representation of the style to a path, and distorted the piecewise polynomial curve or polygon path to create a stylized line. Other approaches reproduced and synthesized similar styles from examples lines [Finkelstein and Salesin, 1994, Jodoin et al, 2002, Kalnins et al, 2002, Freeman et al, 2003, Simhon and Dudek, 2004, Brunn, 2006.…”

Section: Simulating Human Line Drawingsmentioning

confidence: 91%

Automatically mimicking unique hand-drawn pencil lines

AlMeraj

Wyvill

Isenberg

et al. 2009

Computers & Graphics

View full text Add to dashboard Cite

In applications such as architecture, early design sketches containing accurate line drawings often mislead the target audience [Schumann et al., 1996]. Approximate human-drawn sketches are typically accepted as a better way of demonstrating fundamental design concepts. To this end we have designed an algorithm that creates lines that perceptually resemble human-drawn lines. Our algorithm works directly with input point data and a physicallybased mathematical model of human arm movement. Our algorithm generates unique lines of arbitrary length given the end points of a line, without relying on a database of humandrawn lines. We found that an observational analysis obtained through various user studies of human lines made a bigger impact on the algorithm than a statistical analysis. Additional studies have shown that the algorithm produces lines that are perceptually indistinguishable from that of a hand-drawn straight pencil line. A further expansion to the system resulted in mimicked dashed lines.

show abstract

Learning style translation for the lines of a drawing

Cited by 48 publications

References 13 publications

Lighting and Pose Robust Face Sketch Synthesis

Lighting and Pose Robust Face Sketch Synthesis

Stroke Pattern Analysis and Synthesis

Automatically mimicking unique hand-drawn pencil lines

Contact Info

Product

Resources

About