OCR is identifying characters incorrectly

[saikrishnagopal1227]

Current Behavior

After running OCR on an image, I found multiple character recognition errors (e.g., "T" was interpreted as "I"), along with several other misidentified characters.

Expected Behavior

All the characters should identify correctly after OCR.

Suggested Fix

No response

tesseract -v

5.5.1

Operating System

Ubuntu 20.04

Other Operating System

No response

uname -a

No response

Compiler

No response

CPU

No response

Virtualization / Containers

No response

Other Information

Im attaching the file that I have used here

P1.PC.00000003-1 (1).pdf

[amitdo]

Expecting 100% accuracy is not realistic.

[stweil]

This report has missing information. Which model was used? Which image was processed? Attaching a PDF does not help because Tesseract cannot process PDF files.

[saikrishnagopal1227]

I am aware that Tesseract does not process PDF files. My apologies—I accidentally attached a PDF instead of the intended image files. This was an honest oversight. I am now attaching the correct image files that I used for testing.

The language used was English.

Please let me know if any further details are required.

dffce4cf-d4d6-414b-8451-a9f4f02dbc15.tif

1116fdd5-f9a8-44f6-a243-1e15e5fd5d7f.tif

[stweil]

The text of the first image is recognized with tessdata_best/eng.
The 2nd image is more challenging because of the very small text parts, but much of it is recognized well, too.