1Z0-184-25 Question 153
Single answerYour e-commerce platform stores thousands of product images in Oracle Object Storage. The marketing team wants an AI-based solution to automatically classify each image by brand and product category. They also need to add new categories regularly. Which approach best satisfies these requirements using Oracle Cloud Infrastructure� AI capabilities?
- A
Enable Object Storage bucket versioning and rely on built-in tagging capabilities to classify images automatically.
- B
Use Oracle Data Labeling Service to label images by brand and category, train a custom model in Oracle Vision, and periodically retrain the model as new categories are added.
- C
Deploy an on-premises AI platform and regularly push classification updates to objects in Oracle Object Storage.
- D
Configure a scheduled job in Oracle Cloud Shell to rename object files with brand and category information for future reference.
Show answer and explanation
Correct answer: B
Explanation
Leveraging Oracle Data Labeling Service and Oracle Vision is the recommended approach for image classification in OCI. Data Labeling Service provides a straightforward way to tag and manage training data for custom model creation. Oracle Vision� custom models can then be trained on these labeled images to identify product brand and category. Periodic re-labeling and retraining ensure the model evolves to accurately classify newly introduced categories. Refer to Oracle Vision and Data Labeling documentation for best practices on model training and retraining strategies.
- A. Incorrect.
Incorrect. While bucket versioning can track file changes, it does not provide an automated AI-based image classification capability. Built-in object metadata tagging cannot automatically determine brand or category from the image content.
- B. Correct.
Correct. Using Oracle Data Labeling Service to label images is a best practice to build high-quality training data. Oracle Vision can then train custom classification models on these labels. As new categories emerge, you can update the training dataset and retrain the model to keep its predictions accurate.
- C. Incorrect.
Incorrect. An on-premises AI solution would require extensive setup, maintenance, and data transfer overhead. It does not optimally leverage Oracle Cloud Infrastructure� native AI services designed for image classification.
- D. Incorrect.
Incorrect. Renaming files only changes the object name and cannot build or maintain a trained model for AI-based classification. It also doesn�t offer a scalable mechanism to add new categories dynamically.