AI-900 Question 158
Select 3You are a data scientist tasked with building an object detection solution using Azure Machine Learning. Which features of an object detection solution should you consider to ensure it meets the requirements for identifying and localizing objects in images?
- A
Bounding box generation to locate objects in the image
- B
Classification of detected objects into predefined categories
- C
Automatic image enhancement to improve detection accuracy
- D
Integration with Azure Cognitive Search for indexing results
- E
Confidence scoring to indicate the likelihood of correct detection
Show answer and explanation
Correct answers: A, B, E
Explanation
Object detection solutions must be able to locate objects (bounding box generation), classify them into predefined categories, and provide confidence scores to assess the reliability of the results. These features ensure the solution can effectively identify and localize objects in images. Other features, such as image enhancement or integration with unrelated services, are not directly part of core object detection functionality.
- A. Correct.
Bounding box generation is a key feature of object detection as it identifies the location and size of detected objects within an image.
- B. Correct.
Classification is essential for object detection, as it assigns detected objects to specific categories or classes.
- C. Incorrect.
Automatic image enhancement is not a standard feature of object detection solutions. While preprocessing techniques can improve detection, this is not inherently part of the object detection process.
- D. Incorrect.
Integration with Azure Cognitive Search is unrelated to the core functionality of object detection. It is a separate service for indexing and searching content.
- E. Correct.
Confidence scoring is a critical feature that helps determine the reliability of the detected objects and their classifications.