Answer-first summary for fast verification
Answer: Choose an object detection model, label images with bounding boxes and class labels, evaluate using mean average precision.
For detecting multiple objects within images, an object detection model is more suitable as it can identify and locate multiple objects in an image. Labeling involves drawing bounding boxes around each object and assigning class labels. The most relevant metric for evaluating such a model is mean average precision (mAP), which measures the average precision over multiple IoU thresholds.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
In a scenario where you need to detect multiple specific objects within images, such as vehicles in a traffic monitoring system, which type of model would you choose and why? Additionally, explain how you would label the images for this specific use case and what metrics would be most relevant for evaluating the model's performance.
A
Choose an image classification model, label images with object tags, evaluate using accuracy.
B
Choose an object detection model, label images with bounding boxes and class labels, evaluate using mean average precision.
C
Choose an image classification model, label images with bounding boxes, evaluate using precision.
D
Choose an object detection model, label images with object tags, evaluate using recall.
No comments yet.