Extraction is not proper

[Suganya_velu]

the below is the image passed to GCP vision API 

but the expected result it not proper

the second occurrence is not working as expected. this occurs when we give it as PDF format.