Application Note: Creating and Using forms for data capture
Part 1
OMR, OCR and ICR are three terms that one hears frequently in the digital form handling context. In this three part article, I will explore these applications and their practical utilization.
First, let us get the definitions out of the way. OMR is Optical Mark Reading and refers to bubble entry forms (fig 1) which we encounter frequently in high volume survey, application or feedback forms.
OCR is Optical Character reading and uses a software that intelligently maps the printed image of a character to a data point while ICR software uses a far more powerful recognition engine to read handwritten characters….
Part 1
OMR, OCR and ICR are three terms that one hears frequently in the digital form handling context. In this three part article, I will explore these applications and their practical utilization.
First, let us get the definitions out of the way. OMR is Optical Mark Reading and refers to bubble entry forms (fig 1) which we encounter frequently in high volume survey, application or feedback forms.
OCR is Optical Character reading and uses a software that intelligently maps the printed image of a character to a data point while ICR software uses a far more powerful recognition engine to read handwritten characters….
Naturally, OMR forms have the highest accuracy of data trapping , while OCR software have evolved to deliver a digital conversion accuracy of 80-90%+. However, the variations of handwritten script limit the efficacy of the ICR technology to ~60%.
We will review the OMR technology in this issue. OMR hard-wired scanners of yesteryear are now under pressure with the new generation document scanners combining with software like Remark, PC OMR and Creative to deliver plain-paper functionality and power on the small office users’ desks.
How do these software work?
The basics is more or less the same –
1. You create a form template using a word-processor or graphics package
2. You read the printed copy in the scanner and define the basic image/ data areas. In software like Creative ICR, you will define the registration of the page using registration marks. PC OMR , on the other hand, defines standard answer sheets and therefore dispenses with the luxury of custom forms.
3. Finally, you scan the respondents’ copies of the forms – and the software reads the image/data areas as defined in the previous step – and extracts the information into a worksheet – and reports!
How do you design a form in Remark OMR software?
1. Download & install the bubble-script typefaces, called OMR Bubbles, from http://www.principiaproducts.com/
2. Prepare a form using a word-processor like MS Word or a graphics package like Corel Draw.
Following have to be kept in mind while designing the form to get the best possible recognition results. Once you determine the questions you want to ask your respondents, creating the form in a scannable layout is the next step.
Following have to be kept in mind while designing the form to get the best possible recognition results. Once you determine the questions you want to ask your respondents, creating the form in a scannable layout is the next step.
• Mark Selection
o Although Remark Office OMR works with almost any complete shape (circle, square, triangle), experience indicates that bubbles between 10 and 14 points in height work the best. We recommend a capital “O’ in an Arial font for a bubble. Respondents will typically bubbles more completely and neatly. It is possible to place numbers or letters inside of bubbles. However, the number/letter must be as small and light as possible. OMR Bubbles font from the above site gives excellent results…and also installed automatically with the Remark Office software.
o Although Remark Office OMR works with almost any complete shape (circle, square, triangle), experience indicates that bubbles between 10 and 14 points in height work the best. We recommend a capital “O’ in an Arial font for a bubble. Respondents will typically bubbles more completely and neatly. It is possible to place numbers or letters inside of bubbles. However, the number/letter must be as small and light as possible. OMR Bubbles font from the above site gives excellent results…and also installed automatically with the Remark Office software.
• Spacing
o White space is important both between the marks in a field and around the area of the marks. Allow at least 3/8 inch between any text, lines or graphics on the form and any user-markable areas. The more white space the better. Minimise errors by separating marks in a group from one another by at least two character spaces.
o Try to stay away from using lines or boxes around or between the marks on the form. If you find it necessary to use lines or boxes, consider making the lines a light gray (e.g. 5%) , that will drop out (completely disappear) during the scanning process. Use at least ½ inch of space between lines and any user-markable areas of the form. These spacing guidelines allow the software to accommodate skew or offset during the scanning process as well as poorly marked forms.
o White space is important both between the marks in a field and around the area of the marks. Allow at least 3/8 inch between any text, lines or graphics on the form and any user-markable areas. The more white space the better. Minimise errors by separating marks in a group from one another by at least two character spaces.
o Try to stay away from using lines or boxes around or between the marks on the form. If you find it necessary to use lines or boxes, consider making the lines a light gray (e.g. 5%) , that will drop out (completely disappear) during the scanning process. Use at least ½ inch of space between lines and any user-markable areas of the form. These spacing guidelines allow the software to accommodate skew or offset during the scanning process as well as poorly marked forms.
• Grouping questions
o Place similar type of questions together. For example, place all True/False questions in one area of the form and all multiple choice questions in another. Positioning all of the marks for the same question type in a single field area minimizes your efforts in the field definition windows when creating the template. All marks for a single field must align horizontally and vertically. However, the space between rows of bubbles in a group does not have to be equal.
o When you have headings over rows and columns of marks, leave a larger amount of white space between the header row and the first row of marks than you have between the rest of the rows of marks. This will help the Remark software understand where the headings are located and where the actual marks begin. Again – do not place labels between marks – then you may have to select questions individually or join marks together to form logical questions.
o Place similar type of questions together. For example, place all True/False questions in one area of the form and all multiple choice questions in another. Positioning all of the marks for the same question type in a single field area minimizes your efforts in the field definition windows when creating the template. All marks for a single field must align horizontally and vertically. However, the space between rows of bubbles in a group does not have to be equal.
o When you have headings over rows and columns of marks, leave a larger amount of white space between the header row and the first row of marks than you have between the rest of the rows of marks. This will help the Remark software understand where the headings are located and where the actual marks begin. Again – do not place labels between marks – then you may have to select questions individually or join marks together to form logical questions.
• Paper and shading
o Keep it simple ( plain white paper and black ink). Else make sures that coloring drops out when scanned. The paper color must be light enough to disappear so that it does not interfere with the recognition of your optical marks.
o You can raise the brightness setting on your scanner to help shading or color from paper drop out. However, the shading or color on the paper still needs to be light enough so that you do not have to raise the brightness setting too high; this could compromise the rest of the form.
o Keep it simple ( plain white paper and black ink). Else make sures that coloring drops out when scanned. The paper color must be light enough to disappear so that it does not interfere with the recognition of your optical marks.
o You can raise the brightness setting on your scanner to help shading or color from paper drop out. However, the shading or color on the paper still needs to be light enough so that you do not have to raise the brightness setting too high; this could compromise the rest of the form.
• Photocopying forms
o If making photocopies, create your template on the photocopied version rather than the original and do not mix copies from different copiers. This will allow for any offset produced by the photocopier. Change your printing option if there are Toner streaks/ speckles.
o If you plan to use shading on the form, you should not photocopy forms for distribution.
o If making photocopies, create your template on the photocopied version rather than the original and do not mix copies from different copiers. This will allow for any offset produced by the photocopier. Change your printing option if there are Toner streaks/ speckles.
o If you plan to use shading on the form, you should not photocopy forms for distribution.
• Testing
o Always test a form before printing and distributing. This will validate your form design and the template.
Now that the form is ready – just scan it and save it as an image. Next, define your image and data area using the software. Once you save this form template, you are ready to go! Remember to print original forms and use only one paper type instead of making cheap copies and paper from different sources (you could risk registration issues and reading accuracy will inevitably suffer)!
o Always test a form before printing and distributing. This will validate your form design and the template.
Now that the form is ready – just scan it and save it as an image. Next, define your image and data area using the software. Once you save this form template, you are ready to go! Remember to print original forms and use only one paper type instead of making cheap copies and paper from different sources (you could risk registration issues and reading accuracy will inevitably suffer)!
When you read the filled forms from the Remark Software after selecting the form template –you will be pleasantly surprised to see the data getting captured automatically in the process of scanning. You can also generate the reports at one touch using the standard or custom-made reports!
Ashwini Aggarwal is an ex-HP Veteran and consultant on document management and marketing services. He can be contacted at aka@xl-inspire.com for any queries on this article – or for any form review service. For more samples of 'before/after form , ref below - just join as a follower & write in at aka@xl-inspire.com.
Ashwini Aggarwal is an ex-HP Veteran and consultant on document management and marketing services. He can be contacted at aka@xl-inspire.com for any queries on this article – or for any form review service. For more samples of 'before/after form , ref below - just join as a follower & write in at aka@xl-inspire.com.
Sample 1> CPM Event Feedback form (at the start of the article)
Sample 2>Tourist Bureau survey (before with errors) …and (after validation) (write back..)
Sample 2>Tourist Bureau survey (before with errors) …and (after validation) (write back..)
