ChatGPT 能扫描问题吗？

This means for physical books or papers, one needs to use a scanner or phone camera app with OCR before using ChatGPT for Q&A.

ChatGPT itself cannot scan images or physical documents. It processes text input. However, you can use OCR (Optical Character Recognition) tools to convert images or scanned documents into text, then input the text into ChatGPT.

I tried uploading a photo of a handwritten question to ChatGPT with image input and it was surprisingly good at understanding and answering it!

With the latest GPT-4 vision capabilities, ChatGPT can analyze images if you use the right interface, but it still requires the image to be uploaded digitally. It can't scan physical documents directly but can answer questions about an image you provide.

So in summary, ChatGPT doesn't 'scan' questions by itself but can understand text extracted from images or documents if given the text or image input in the right format.

Does the OCR accuracy affect how well ChatGPT can answer questions from scanned images?