Interpreting the density string values
By using the density string characters and the corresponding formula, you can calculate the percentage of black pixels for each OMR zone.
The previous examples produce these results.
| Parent field | Subfield | Checkbox | Density value | Percentage filled |
|---|---|---|---|---|
| Options | Nav_System | Cleared | A | 17% |
| Child_Seat | Cleared | C | 19% | |
| Fuel_Service | Checked | G | 23% | |
| Insurance | CDW | Cleared | C | 19% |
| PAI | Cleared | C | 19% | |
| PEP | Cleared | B | 18% | |
| ELP | Checked | G | 23% |
By using the data in first table, you can determine that threshold = 20.5 and background = 20 represent good values. The (2 * threshold - background) formula sets the high-confidence threshold at 21, but you can scan more pages and check their density strings before you set final values.
| Parent field | Subfield | Checkbox | Percentage filled | Result in runtime hierarchy |
|---|---|---|---|---|
| Options | Nav_System | Cleared | 17% | Option: Selected Confidence: High |
| Child_Seat | Cleared | 19% | Option: Not selected Confidence: High |
|
| Fuel_Service | Checked | 23% | Option: Selected Confidence: High |
|
| Insurance | CDW | Cleared | 19% | Option: Selected Confidence: High |
| PAI | Cleared | 19% | Option: Not selected Confidence: High | |
| PEP | Cleared | 18% | Option: Not selected Confidence: High | |
| ELP | Checked | 23% | Option: Not selected Confidence: High |
Because the OMR zones that you drew are probably different from the ones that generated the data in second table, your values are different. Use the density characters that you obtained to determine appropriate values for the threshold and background parameters, and update the Recognize OMR Fields rule. Then, run a new batch and review the results, including the confidence values in the page data files.