As the name suggests, mask-words-in-image can mask words in an image file. User specifies which words should be masked, it can be any combinations of the following three types.
- Regular expressions
- Keywords
- PII (Personally Identifiable Information)
Under the hood, it use the Text Detection capability in Amazon Rekognition or AWS Textract to detect the texts in the image, then use Amazon Comprehend to detect PII entities in texts.
Here is a quick demo:
Jackie Chen's IT Workshop 