Minimum Font Size, DPI, and Image Quality Metrics for Optimal Google Vision Accuracy

[hahh]

Hi everyone,

I’m currently working on a project that relies heavily on Google Vision’s OCR capabilities, and I want to ensure that the images we process are optimized for maximum accuracy. I’ve reviewed the Google Vision API quotas and limits(https://cloud.google.com/vision/quotas#limits), but I couldn’t find specific details about the optimal image quality parameters.

Could anyone provide insights or recommendations on the following aspects to improve text detection accuracy?

1. Minimum font size: What is the smallest font size that Google Vision can reliably detect in images?

2. DPI (Dots Per Inch): What is the recommended DPI for images to ensure high OCR accuracy?

3. Other image quality metrics: Are there any other factors (e.g., contrast, resolution, noise levels) that significantly impact Google Vision’s ability to accurately detect and process text?

I appreciate any guidance or experiences you can share regarding these parameters to help fine-tune our image preprocessing pipeline.