After the first 50 ballots scanned with the pyhton iOCR code, it was determined the program was 80% accurate. I was tasked to take my code and to reorganize it to make it more acccurate with scanning. Taking it out meaningless codes and replace it with more efficient code. Then to scan more ballots with the new code to hopfully increase the accuracy of results.

I was tasked to create the python verision of the code that will read in a file, cconvert the (.png or text.) and use levenshtien distance, to create a new iOCR Json file. Then to send the Json to display to HTML , without using a local server. This is completed and when we used the code in the lab it became more difficult since it was used in previous device. Still adjusting the code and it is coming along really well.

Once I have did my test run with the scanner that yeild the best results we configured an plan to use a different Optical Character Recognition(OCR) than the one built-in the scanner. we decided tesseract to be out other OCR, and used the scanner to test out how accurate its OCR was in different font sizes from both scanners. I was then tasked to program code to do thesame thing the built in scanner is doing but in python, then once that was complete I was than tasked to compare the txt.file scan to the ballot definition using Levenshtein Distance in developments of iOCR (Intelligent Optical Character Recognition). still working on the Levenshtein distance to be alot more accurate with the write-ins on the ballots.

We met and discussed the different direction we want the application to function if it could be on windows or if it could be stateless, using either python or JavaScript.Then I was tasked to take a regular ballot and convert it to Html format. once completed I was instructed to read a couple of papers like: “RISK MANAGEMENT FOR ELECTRONIC BALLOT DELIVERY, MARKING, AND RETURN”, “A Study of Ballot Anomaly Detection with a Transparent Voting Machine”. After that I was given a brief demo of how it works. Fianlly i was tasked to use the scanners to see which one is easier to navigate when scanning ballots.

When I started the project I was in countless meetings. Then I was tasked to find particular scanners that were: Linux and Windows compatible, Included the OCR software, and able to be mobile. after completing that , I was tasked to research “what makes a Webpage Screen reader friendly ?” , and “what are the best practices of designing a webpage that is screen reader friendly ?”, and to learn about “tesseract”. finally was to read about Risk Management for Electronic Ballot Delivery, Marking, and Return.