Impressive results. From the three samples provided, the original image quality is not super high but the program was still able to figure out the characters for all but a very few.
For Sample1:
All text detected accurately.
For Sample2:
varicties vs. varieties
Scction vs. Section
usc vs. use
italso vs. it also
ITat vs. III at
For Sample 3:
MarGARET vs. MARGARET (small caps, so there is a )
quarrelled vs. quarreled
cleven vs. eleven
occurrenceseemsclear vs. occurrence seems clear
ccstatically vs. ecstatically