Image Processing and OCR

Question 1

What does OCR stand for

Answer 1

Optical Character Recognition

Question 2

OCR turns text into?

Answer 2

image-based content into machine-readable text

Question 3

What are the 3 OCR Engines that come with all Grooper installs

Answer 3

Tesseract OCR
Transym 4 OCR
Transym 5 OCR

Question 4

Matrix matching and feature recognition are part of what phase of an OCR engine’s operation?

Answer 4

Character Recognition

Question 5

Breaking up pixels into lines, words, and characters is part of what phase of an OCR engine’s operation?

Answer 5

Segmenting

Question 6

Many OCR engines spell check OCR results to improve
their accuracy. This is part of what phase of an OCR
engine’s operation?

Answer 6

Post-Processing

Question 7

In your own words, describe the Segmenting phase of an

OCR engine.

Answer 7

This is when the pixels are broken up into lines, individual words, and
characters

Question 8

OCR engines that obtain results by comparing a grid of
pixels on an image to a grid of pixels of examples of
characters are performing….

Answer 8

Matrix Matching

Question 9

The Grooper activity that performs OCR is….

Answer 9

Recognize

Question 10

What image processing operation is required for an OCR
engine to obtain results, either through Grooper’s image
processing suite or via the OCR engine itself?

Answer 10

thresholding (or binarizing) the image

Question 11

Image processing in Grooper serves one (1) of three (3)

basic purposes. What are they?

Answer 11

Archival Adjustments also OCR Cleanup and Layout Data collection (Archival Adjustments
ONLY pertain to permanent image processing via the Image
Processing activity)

Question 12

The Grooper activity that performs permanent image

processing is….

Answer 12

Image Processing

Question 13

In your own words, what is the benefit of performing
temporary image processing? How do you perform
temporary image processing in Grooper?

Answer 13

Temporary Image Processing is great because it will not make permanent
changes to the document itself. You assign a Temp IP Profile and run the
recognize activity. The only thing I will add is where that temporary IP
Profile gets assigned. It is assigned on the OCR Profile (which then
gets executed by the Recognize activity).

Question 14

List three (3) common IP Commands used during
permanent image processing.

Answer 14

Auto Deskew, Auto Border Crop, Rotate

Question 15

List three (3) common IP Commands used during
temporary image processing.

Answer 15

Line Removal, Speck Removal, Negative Region Removal

Question 16

Grooper’s set of properties that pre-process and reprocess

the OCR engine’s results are called…

Answer 16

Synthesis

Question 17

There are five (5) operations that comprise this

synthesis functionality. List them

Answer 17

Font Pitch Detection, Bound Region Processing, Iterative Processing, Cell
Validation, Segment Reprocessing

Question 18

Where are the synthesis properties enabled and configured?

Answer 18

ON an OCR Profile

Question 19

In your own words, what is “fuzzy regular expression”?
How does “fuzzy regular expression” improve Grooper’s
ability to extract data from poorly OCR’d pages?

Answer 19

This allows you to match expressions and set a percent match of how
close it looks like what you are trying to find, this helps eliminate errors
when extracting data.

Question 20

How do you alter the normal cost to swap characters

when using fuzzy regular expression?

Answer 20

Fuzzy Match Weightings

Question 21

How do you force a portion of a fuzzy regular expression

to match normally (or non-fuzzily)?

Answer 21

Required Mode

Question 22

Non-text information obtained via permanent or
temporary image processing such as line locations,
checkbox locations and states, barcode values, and
detection of trained shapes is referred to as…

Answer 22

Layout Data

Question 23

Once Non-text information obtained via permanent or
temporary image processing such as line locations,
checkbox locations and states, barcode values, and
detection of trained shapes is collected, where is this information stored in Grooper?

Answer 23

The LayoutData.json file. It is stored on each page object a Layout Data IP Command locates non-text data.

Question 24

What “tab marking” property will insert a tab character
(“\t”) between the highlighted values in the table below,
without adjusting the width of a tabbed space?

Answer 24

Detect Lines

Question 25

True or False: For the previous question, the lines must
be detected by a Line Removal or Line Detection
command BEFORE the extractor executes in order to insert
the tab character.

Answer 25

True