You are developing a model to identify traffic signs in images extracted from videos taken from the dashboard of a vehicle. You have a dataset of 100,000 images that were cropped to show one out of te

This is an image classification problem, not object detection, because the images are already cropped to show a single object. Vertex AI AutoML would also offer less control for tuning than custom training.

While it's an image classification problem, Vertex AI AutoML generally provides less fine-grained control for model tuning during *each training run* compared to custom training.

This is an image classification problem, not object detection, as the objects are already isolated and cropped within each image.

Since the images are *cropped to show one* of ten different traffic signs, this is an image classification problem, not object detection. Developing custom training code and using Vertex AI Custom Training allows full control over the model architecture and hyperparameter tuning during each training run, meeting the tuning requirement.