AI-900 Question 146
Select 3A retail company wants to use computer vision to enhance their store operations. They are considering solutions to automatically detect when shelves are empty, recognize specific products, and analyze customer behaviors like heatmaps of foot traffic. Which types of computer vision solutions would be most relevant for their needs?
- A
Object detection
- B
Image classification
- C
Optical character recognition (OCR)
- D
Semantic segmentation
- E
Crowd analysis
Show answer and explanation
Correct answers: A, B, E
Explanation
The tasks described, detecting empty shelves, recognizing products, and analyzing customer behavior, align with object detection, image classification, and crowd analysis. Object detection helps locate specific products or empty shelves, image classification categorizes visual data like shelf status, and crowd analysis provides insights into customer movements and behaviors. OCR and semantic segmentation are not directly applicable to the given scenario.
- A. Correct.
Object detection is suitable for identifying and locating specific items, such as recognizing empty shelves or detecting products in an image.
- B. Correct.
Image classification can be used to categorize images, such as identifying whether a shelf is empty or full based on a predefined set of labels.
- C. Incorrect.
OCR is used for extracting text from images, such as reading labels or printed text, which is not directly related to the scenario described.
- D. Incorrect.
Semantic segmentation is used for pixel-level understanding of an image, which is more suited for tasks like autonomous driving rather than detecting shelves or analyzing customer behaviors.
- E. Correct.
Crowd analysis, a type of computer vision solution, is relevant for generating heatmaps of foot traffic and understanding customer behavior patterns in the retail store.