AI-900 Question 20
Single answerA retail company wants to implement a solution to automatically identify products on their store shelves using images captured by cameras. Which type of computer vision workload would be the most appropriate for this scenario?
- A
Image classification
- B
Object detection
- C
Optical character recognition (OCR)
- D
Semantic segmentation
Show answer and explanation
Correct answer: B
Explanation
Object detection is the most appropriate workload for this scenario because it can identify specific objects (products) within an image and determine their locations. This functionality is essential for the retail company to analyze images of store shelves and recognize individual products effectively.
- A. Incorrect.
Image classification assigns a single label to an entire image, but it does not identify specific objects or their locations within the image. This does not meet the requirement to detect individual products.
- B. Correct.
Object detection identifies and locates multiple objects within an image, making it well-suited for detecting specific products on store shelves.
- C. Incorrect.
Optical character recognition (OCR) is used to extract text from images, such as reading labels or receipts. This is not applicable for identifying and locating products.
- D. Incorrect.
Semantic segmentation classifies each pixel in an image into a category, which is useful for understanding image regions but is not the most efficient approach for identifying and locating individual products.