Microsoft Azure AI Fundamentals (AI-900) Practice Exam

To determine the exact boundaries of a bear in an image, which technique should be applied?

1. Optical character recognition (OCR)

2. Image classification

3. Object detection

4. Semantic segmentation

The correct answer is: Semantic segmentation

Semantic segmentation is the most appropriate technique for determining the exact boundaries of a bear in an image. This approach involves classifying each pixel in the image to a particular category, thereby providing a detailed understanding of the shapes and contours of objects within that image. By applying semantic segmentation, you can delineate the bear from the background as well as accurately capture its shape, providing pixel-level precision. In contrast, image classification identifies what object is present in an image but does not provide any information about the boundaries or location of that object within the image. Object detection identifies and localizes multiple objects in an image with bounding boxes but does not provide the detailed boundary information necessary to outline the bear precisely. Optical character recognition (OCR) is geared towards identifying text in images, making it irrelevant for tasks involving identifying and outlining objects like a bear.