Data Set of Drawings

by Yutaka Ouchi and Kazunori Iwata last modified Jan. 12, 2012

1. What is the Data Set ?

The data set consists of 100 text format files, each of which represents a letter, mark, etc. Each line of the file is expressed as a vector (t, x(t), y(t)) where t denotes time (milli-second) which has passed from the start of drawing line, and x(t) and y(t) denote the coordinates of point at time t.

2. How to get ?

The data set is free and everyone is permitted to use them. We will accept that you have the freedom to distribute the data set. Before downloading, please understand that we abdicate responsibility for any trouble and any technical support regarding the data set. You must accept this statement to download it.

Download

dataset.ZIP

3. How to use ?

You can visualize a text file of the data set using gnuplot, Drawing Panel, MS Excel, etc.

With gnuplot

When plotting a file "xxx.txt" with gnuplot, launch as follows.

> set yrange [] reverse
> plot "xxx.txt" using 2:3 with lines

With Drawing Panel

Selecting ``Open'' item will plot a file.

References

K. Ueaoki, K. Iwata, N. Suematsu, and A. Hayashi, Matching Handwritten Line Drawings with Von Mises Distributions, IEICE Trans. Info. and Syst., Vol. E94-D, No. 12, pp. 2487-2494, 2011.

Y. Ouchi, K. Iwata, N. Suematsu, A. Hayashi, Representing with Kernel Functions to Recognize Shapes, IEICE Trans. Info. and Syst., Vol. J92-D, No. 11, pp. 2011-2021, 2009. (in Japanese)

[B A C K]