OpenSketch: A Richly-Annotated Dataset of Product Design Sketches

Yulia Gryaditskaya Université Côte d'Azur, Inria
Mark Sypesteyn Delft University of Technology, Faculty of Industrial Design Engineering
Jan Willem Hoftijzer Delft University of Technology, Faculty of Industrial Design Engineering
Sylvia Pont Delft University of Technology, Perceptual Intelligence lab
Frédo Durand MIT CSAIL, Inria
Adrien Bousseau Université Côte d'Azur, Inria

SIGGRAPH Asia 2019

[Paper]

Stroke-processing library and I/O of OpenSketch format: [pyLowStroke]

Sketch-reconstruction library: [pySBM]

Content

Use the menu on the left to navigate through the dataset:
Collected sketches - all the collected sketches.
Annotations: 3D shape Registration - all the collected sketches overlayed with registered 3D models.
Annotations: Lines labeling - each line type is shown in a separate layer for all the sketches made from the first viewpoint (paper: Section 6.1 Stroke labeling).
Applications: lines classification - visualization of automatic clustering of lines for two groups.
Applications: concept to presentation - results of concept to presentation conversions on test data.
Applications: sketch to normal - results of normals prediction from the input sketches.
The images are gouped either by a designer (Grouped by designer) or by an object (Grouped by object).

All downloads

For details, see downloads pages from the menu on the left.

Downloads:

All above as a single zip file: Additonal files:

Code:

Additional files

It additionally contains:
supplemental.pdfThe document containing the information complementary to the paper, displayed bellow
example_instructions.pdf The instructions which the participants were provide prior to sketching as well as a screenshot of the drawing interface.
strokes_types_distribution.pdf Pie-charts computed per each object within students and professionals groups
drawing_explorer.py The simple python viewer to explore the drawing creation process.

Supplemental pdf file

Example instructions