Browse Prior Art Database

Method and software for extracting numerical values of curves from digitized images

IP.com Disclosure Number: IPCOM000216237D
Publication Date: 2012-Mar-26
Document File: 7 page(s) / 92K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method and software implementation for recovering/extracting precise numerical values from a digitized figure.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 31% of the total text.

Page 01 of 7

Method and software for extracting numerical values of curves from digitized images

For reference, comparison, or processing purposes, analysts often need to extract the precise numerical values that correspond to data points plotted in a chart within a published document (e.g., on the Internet, in PDF documents, articles, etc.). In that format, however, the data is a digitized figure and is not available as numerical values.

The disclosed solution is a method and software implementation in MATLAB* for recovering/extracting the precise numerical values from the digitized figure.

The new method consists of the following steps: (Fig. 1)


1. Loading digitized image

2. Display digitized image into a tool for extracting the precise numerical value of the figure data


3. Define X and Y axis, range limits, scale using a pointing device


A. Define a starting and ending point of the X axis on the displayed figure

B. Enter a numerical value for both the starting and ending point of the identified X axis


C. Define a starting and ending point of the Y axis on the displayed figure

D. Enter a numerical value for both the starting and ending point of the identified Y axis


E. Specify for box X and Y axis if they are linear axis or logarithmic axis

F. Possibly, display the numerical axis onto the digitized figure for visual confirmation of the operation

4. Identify points along the curve in the digitized figure using a pointing device such as a mouse

5. Repeat the previous operation until the digitized curve has been replicated with sufficient accuracy. For each new point, numerical points may be plotted on the digitized picture to show partial results, the points may also connect with a curve.

6. Calculate precise numerical values of the points along the curve in the digitized picture based on the defined axis (step 3) and relative locations of the selected points (steps 4 & 5)

7. Save the calculated numerical values into a file, such as a text file, numerical data file, excel file, etc and/or display the data as a numerical plot curve

Figure 1: Steps for implementation

1


Page 02 of 7

Referring to the method described in Fig. 1, a particular embodiment based on MATLAB software follows:

From the MATLAB command shell the "curve_extract.m" script is executed. During a first step a dialog box (Fig. 3) is opened to allow the user to select and open a digitized figure (Fig. 2) file.

Subsequently, a MATLAB figure window (Fig. 4) is created to show the loaded digitized figure (Fig. 2) to the user.

Figure 2: Example of digitized picture

Figure 3: Load dialog box

2


Page 03 of 7

Figure 4: Analysis window

The software requests the user to identify the axis by using a pointing device such as a mouse-controlled cursor. During this process, the user may be allowed to zoom into the figure to magnify specific regions of the digitize image for better identifying reference points. In one embodiment, the user has to click in four different locations identifying the be...