SIGGRAPH 2010 Structurebased ASCII Art Xuemiao Xu Linling

SIGGRAPH 2010 Structure-based ASCII Art Xuemiao Xu, Linling Zhang, Tien-Tsin Wong The Chinese University of Hong Kong

Since the 1860 s, text art emerged…

From the 1970 s, ASCII art has been widely used…

Today, ASCII art remains popular…

ASCII Art Classification • Tone-based – Halftone approaches Regarded as dithering – O’Grady and Rickard [2008] Dithering essentially • Structure-based

ASCII Art Classification • Tone-based • Structure-based – Halftone approaches – Manual Regarded as dithering Tedious Automatic generation of structure-based ASCII art – O’Grady and Rickard [2008] Dithering essentially

Main Challenge • Arbitrary image content

Main Challenge • Arbitrary image content • • Extremely limited character shapes Restrictive placement of characters

Matching Strategies _ _) • Character matching – Misalignment tolerance – Transformation awareness

Matching Strategies _ _) (_ (_ • Character matching – Misalignment tolerance – Transformation awareness • Image deformation – Increase the chance of matching – Avoid over-deformation

Matching Strategies • Character matching Alignment-insensitive – Misalignment tolerance shape similarity metric – Transformation awareness • Image deformation Constrained deformation – increase the chance of matching – Avoid over-deformation

Framework Current best matched Matching error map Input Vectorized Rasterized polylines image characters

Framework _)^ ; r; Good matching Poor matching Current best matched characters Matching error map

Framework (') (_) Current best matched characters Matching error map

Framework (_) Current best cost matched Deformation map characters Matching error Combined cost map

Framework Current bestimage matched Deformed characters Optimal ASCII art Combined cost map

Objective Function • Shape dissimilarity between ASCII and deformed images • Deformation cost of the vectorized images E = DAISS. Ddeform

Main Contribution • Shape Matching Alignment-Insensitive Shape Similarity (AISS) Metric • Constrained Deformation Metric

AISS OCR • Matching requirements • Misalignment tolerance • Transformation awareness • Scope • Pattern recognition and image analysis, e. g. OCR O O 69

Design of AISS • Misalignment tolerance log-polar diagram Log-polar histogram Log-polar diagram (5 x 12)

Design of AISS • Transformation awareness New sampling layout h

Metrics Comparison (1) • Transformation-invariant metrics Query Shape Context Translation and scale invariant Our metric

Metrics Comparison (2) • Alignment-sensitive metrics Query SSIM RMSE after blurring Over-emphasize overlapping Our metric

Main Contribution • Shape Matching Alignment-Insensitive Shape Similarity (AISS) Metric • Constrained Deformation Metric

Constrained Deformation • Local deformation constraint • Accessibility constraint

Local Deformation Constraint A r’ r A’ B’ B

Accessibility Constraint

Optimization Vectorized Input image Corresponding ASCII art

Comparison Input O’Grady & Rickard Our method Resolution=30 X 20 Resolution=20 X 15

User Study Test set 2: 3: Test set 1: Similarity Artists Our method 6. 86 7. 36 O’Grady & Rickard 4. 42 Clarity Input Artists 7. 18 Our method 7. 09 O’Grady & Rickard 4. 15 By Artist Our Method O’Grady & & Rickard

More Results

Other Results

Other Results

Conclusion • Mimic ASCII artists’ work by an optimization process • Propose a novel alignment-insensitive shape similarity metric - also benefits pattern recognition • Propose a new deformation metric to control over-deformation

Limitation • Do not consider the stylish variation of line thickness within a font • Do not handle proportional placement of characters A A • Affected by the quality of the vectorization

Q&A