OCR-ing the First 25

Selecting an OCR Engine

It will take a little over a semester for all 200 cookbooks to be scanned, so I decided to take the first 25 books scanned so far and test out my OCR and text pre-processing steps on them. This will allow me to see whether I have correctly planned the steps I’ll need to take with the full corpus to get meaningful results. Going into the project I had experience with and access to Tesseract OCR, an open-source commend line tool for converting images to machine-readable text, so I decided I would use that to create the .txt files I’ll need for analysis. When I first decided to use Tesseract I had tested a few scanned cookbooks and noticed that Tesseract struggled to recognize the fractions in ingredient lists, but otherwise seemed to do a pretty accurate job of handling the rest of the text.

My colleague in Preservation who is managing digitization is also OCR-ing the texts as part of the digitization workflow we are testing for all digital collections. She is using ABBYY Finereader, a paid license product with a GUI. Since we’re using both products for different parts of the workflow and we have 25 completed, I now have more examples to compare to see if there is a significant difference in accuracy. With more examples to compare, I can now see a noticeable difference in how Tesseract and Finereader handle the table of contents and other non-standard text layouts as well as images and other “noise” on the page.

Determining Pre-Processing Steps

No matter which OCR option I choose, there will inevitably be mistakes that produce extra punctuation, numerical characters, or unrecognizable characters in the text that will need to be cleaned before or ignored by Python during analysis. I’ll be testing different options to determine how much advanced cleanup will be necessary to make sure these errors and extra characters don’t muddy the analysis.

First 25 Cookbooks List

Call Number	Title	Organization	Location	Date
TX715.2 M53 B545x	“Blessed are those who hunger … for they will be filled”	First Congregational Church Women’s Fellowship	Forest City, IA
TX715.2 M53 B6638x 1960	A book of favorite recipes	First United Presbyterian Church	Fairfield, IA	1960
TX715.2 M53 C6514x 1923	A collection of choice and tried recipes	Philathea Society of the Whiting Congregational Church	Whiting, IA	1923
TX715.2 M53 C6515x 1903	A collection of choice recipes	Ladies of Des Moines to benefit Des Moines Missionary Sewing School	Des Moines, IA	1903
TX715.2 M53 C6622x 1950	Cook book	Woman’s Society of Christian Service of the Methodist Church	Blakesburg, IA	1950
TX715.2 M53 F365x 1928	The Farm Bureau cook book	Women of Delaware County Farm Bureau	Delaware County, IA	1928
TX715.2 M53 F38 1960z	Favorite recipes	Altar & Rosary Society, St. Mary’s Catholic Church	Manchester, IA	1960
TX715.2 M53 F753x 1992	Friends of the Fairfield Public Library cookbook	Friends of the Fairfield Public Library	Fairfield, IA	1992
TX715.2 M53 H6622x	Home town recipes of Oskaloosa	Altrusa Club of Oskaloosa	Oskaloosa, IA
TX715.2 M53 K586x 1976	“Kitchen secrets” with love	Low-Rent Housing Agency of Ottumwa, Iowa	Ottumwa, IA	1976
TX715.2 M53 M385x 1896	Mason City cook book: a collection of choice recipes	Woman’s Society of the Congregational Church	Mason City, IA	1896
TX715.2 M53 O67x 1975	Oppskrifter	Palestine Lutheran Church Women	Huxley, IA	1975
TX715.2 M53 O88x 1977	Ottumwa cook book		Ottumwa, IA	1977
TX715.2 M53 O97x 1982	Our best recipes, 1857-1982	Winthrop Commercial Club	Winthrop, IA	1982
TX715.2 M53 R432x	Recipe roundup	American Legion Auxiliary, williams-Jobe-Gibson Unit 128	Sidney, IA
TX715.2 M53 R43x 1971	Recipe favorties of Holt mothers	Des Moines Holtap Mother’s Club	Des Moines, IA	1971
TX715.2 M53 S363x 1999	Schaller Chamber cookbook: celebratin 50 years of pop corn days	Schaller Chamber of Commerce	Schaller, IA	1999
TX715.2 M53 S713x 1910	St. John’s Lutherna Church cookbook	Lutheran Ladies Aid Society	Ogden, IA	1910
TX715.2 M53 T472x 1924	Tested cooky recipes	McGregor Tourist Club	McGregor, IA	1924
TX715.2 M53 T476x 1935	Tested recipes	Orphans’ Aid Society of the St. Paul’s Evangelical Lutheran Church	Waverly, IA	1935
TX715.2 M53 U55x 1997	Unitarian Universalist Fellowship of Ames 50th anniversary cookbook	Unitarian Universalist Fellowship of Ames	Ames, IA	1997
TX715.2 M53 W355x 1970	Walnut centennial cook book	Walnut Centennial Cookbook Committee	Walnut, IA	1970
TX715.2 M53 W54x 1936	When do we eat?	Division One of Willing Workers, Frankville Cmomunity Church	Winneshiek County, IA	1936
TX715.2 M53 W555x 1972	Wild rose recipes	21st National Square Dance Convention	Des Moines, IA	1972
TX723.5 C9 F38x 1963	Favorite recipes of Z.C.B.J. Drill Team	Z.C.B.J. Drill Team	Cedar Rapids, IA	1963