Missing OCR text for many images

[danvk]

A few examples:

• 702198b — text on brown backing paper
• 706410b — brown backing with text
• 709457b — grayscale with text but no OCR
• 729236b — grayscale with text but no OCR
• 711642b — Missing text from color image
• 711564b — Missing text from color image
• 716490b — Missing text from color image
• 731966b — Missing text from gray image (why?)
• 703429b — Missing text from color image

[danvk]

Based on my survey, ~20% of images have text on the back that was not OCR'd.

[danvk]

In that list, 8/9 were missing from the NYPL's S3 bucket. 731966b was actually the front of the image.

@riordan There were 30,413 back of the card images in the S3 bucket, but ~43,000 photos in the CSV file that Matt originally sent me. Is there any chance we could recover more of them?

[riordan]

I'll try.