Assisting people with understanding charts

1. A computer-implemented method comprising:

selecting a chart;

interpreting contents within the selected chart by at least;

recognizing, by the processor, a type of the selected chart, wherein recognizing a type of the selected chart comprises extracting one or more visual features into a learning classifier that provides a class of chart with the highest likelihood of being the selected chart;

fitting, by the processor, data to determine data that represents the contents within the selected chart; and

extracting, by the processor, information from the selected chart and matching the extracted information with the determined data;

searching, by a processor, a plurality of websites for additional sources of information;

extracting, by the processor, additional information from results of the searching of the plurality of websites for the additional sources of information;

combining, by the processor, the interpreted chart contents with the extracted additional information, wherein combining includes weighting the extracted additional information in terms of relevancy to the interpreted contents of the selected chart; and

generating, by the processor, a textual description of the selected chart based on the weighting of the extracted additional information in terms of relevancy to the interpreted contents of the selected chart.

2. The computer-implemented method of claim 1 wherein fitting data to determine data that represents the contents within the selected chart comprises a mean square error method.

3. The computer-implemented method of claim 1 wherein extracting information from the selected chart comprises an optical character recognition method.

4. The computer-implemented method of claim 1 wherein selecting a chart is performed automatically or by a user.

5. A system comprising:

a processor in communication with one or more types of memory, the processor configured to:

select a chart;

interpret contents within the selected chart by at least;

recognizing, by the processor, a type of the selected chart, wherein recognizing a type of the selected chart comprises extracting one or more visual features of the selected chart and inputting the one or more extracted visual features into a learning classifier that provides a class of chart with the highest likelihood of being the selected unit;

fitting, by the processor, data to determine data that represents the contents within the selected chart; and

extracting, by the processor, information from the selected chart and matching the extracted information with the determined data;

search a plurality of websites for additional sources of information;

extract additional information from results of the searching of the plurality of websites for the additional sources of information;

combine the interpreted chart contents with the extracted additional information, wherein combining includes weighting the extracted additional information in terms of relevancy to the interpreted contents of the selected chart; and

generate a textual description of the selected chart based on the weighting of the extracted additional information in terms of relevancy to the interpreted contents of the selected chart.

6. The system of claim 5 wherein the processor being configured to fit data to determine data that represents the contents within the selected chart further comprises the processor configured to execute a mean square error method.

7. The system of claim 5 wherein the processor being configured to extract information from the selected chart further comprises the processor configured to execute an optical character recognition method.

8. The system of claim 5 wherein the processor being configured to select a chart is performed automatically.

9. A computer program product comprising:

a non-transitory storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising:

selecting a chart;

interpreting contents within the selected chart by at least:

recognizing a type of the selected chart, wherein recognizing a type of the selected chart comprises extracting one or more visual features of the selected chart and inputting the one or more visual features of the selected chart and inputting the one or more extracted visual features into a learning classifier that provides a class of chart with the highest likelihood of being the selected chart;

fitting data to determine data that represents the contents within the selected chart; and

extracting information from the selected chart and matching the extracted information with the determined data;

searching a plurality of websites for additional sources of information;

extracting additional information from results of the searching of the plurality of websites for the additional sources of information;

combining the interpreted chart contents with the extracted additional information wherein combining includes weighting the extracted additional information in terms of relevancy to the interpreted contents of the selected chart; and

generating a textual description of the selected chart based on the weighting of the extracted additional information in terms of relevancy to the interpreted contents of the selected chart.

10. The computer program product of claim 9 wherein fitting data to determine data that represents the contents within the selected chart comprises a mean square error method.

11. The computer program product of claim 9 wherein extracting information from the selected chart comprises an optical character recognition method.