Exam AI-900 topic 1 question 243 discussion

B is correct. Images -> Object Detection (Where is the object) -> Image Classification (Get label) -> OCR (Identify the name of place)

Identify street name -> OCR

To identify street signs you should use object detection. OCR is the second step.

To identify street names based on street signs in photographs, you should use optical character recognition (OCR). OCR technology enables computers to recognize and extract text from images, making it suitable for reading and interpreting text on street signs.

according to copilot answer is ocr

answer is B, needs to extract text from signs

Key word is Street Names so it is Image Classifications

Answer is A

Most people have voted for OCR, but the signs are not always text.

B. optical character recognition (OCR) OCR is specifically designed to identify and read printed or handwritten text within images, which is necessary for extracting street names from photographs of street signs. Incorrect Options: A. Object detection: While this could identify the presence of a street sign in an image, it wouldn't read the text on the sign. C. Image classification: This would categorize the image but not provide the detailed text information from the street signs. D. Facial recognition: This is used for identifying and verifying individuals from images or video frames, not for text extraction.

Reading text from images is OCR

OCR - to read the text from the signs. Don't need object detection for this.

too many wrong answers for an AI-900 test preparation, like this one

Answer B - OCR

it's an OCR problem

B. optical character recognition (OCR) To identify street names based on street signs in photographs, you should use **optical character recognition (OCR)**. OCR technology is designed to recognize and extract text from images, making it the most appropriate choice for extracting text content from street signs. Options A, C, and D are not specifically designed for text extraction and analysis but rather for object recognition, image classification, and facial recognition, respectively.

B-OCR, because you are extracting test from images by identifying the characters.