Which computer vision technique would you use to detect multiple objects in a given image?

The technique designed specifically for detecting multiple objects within an image is object detection. This approach not only identifies the presence of various objects but also localizes them by providing bounding boxes around each detected object. Object detection generates multiple labels corresponding to different classes within the same image, making it suitable for applications such as surveillance, autonomous vehicles, and image analysis.

In contrast, optical character recognition (OCR) is focused on recognizing and extracting text from images rather than identifying various distinct objects. Image classification categorizes an entire image into a single label, meaning it is not capable of identifying multiple objects within the same scene individually. Semantic segmentation is concerned with partitioning the image into segments corresponding to different classes, but it does not specifically localize or provide bounding boxes around objects, making it less effective for identifying multiple items simultaneously. Thus, when the requirement is to detect and localize multiple objects, object detection is the most appropriate method.