AI-102 Implement computer vision solutions — Timed 10 Questions

Question 1

You are designing a computer vision solution that must detect and redact personally identifiable information (PII) from images of ID cards. The solution must comply with GDPR. Which THREE components should you include? (Select three.)

10 minute time limit — choose an answer to begin.

Accepted Answer

Azure AI Document Intelligence redaction. Correct options: A, C, D. Azure AI Vision OCR (C) extracts text from ID card images. Azure AI Language PII detection (D) identifies PII entities in the extracted text. Azure AI Document Intelligence redaction (A) can redact the identified PII from the document. Option B (Azure AI Custom Vision) is used for custom image classification, not for text-based PII detection. Option E (Azure AI Video Indexer) is for analyzing video content, not static ID cards.

Answer

Azure AI Custom Vision

Answer

Azure AI Video Indexer

Question 2

You need to analyze a video stream from a security camera to count the number of people entering a building. Which Azure AI service is most suitable?

Accepted Answer

Azure AI Spatial Analysis. Azure AI Spatial Analysis is the most suitable service because it is purpose-built for real-time spatial analysis of video streams, including counting people entering a building. It can process live camera feeds and detect when people cross a line or enter a zone, providing real-time counts. In contrast, Azure AI Video Indexer is designed for analyzing recorded video files, not live streams, making it less suitable for real-time security camera analysis. Other options like Custom Vision and Computer Vision are for image classification and image analysis, respectively, not optimized for video streams.

Answer

Azure AI Custom Vision

Answer

Azure AI Computer Vision

Answer

Azure AI Video Indexer

Question 3

Your retail company uses Azure AI Face API to identify VIP customers when they enter a store. Due to privacy regulations, you must obtain explicit consent and store minimal biometric data. Which approach meets these requirements?

Accepted Answer

Create a PersonGroup with each VIP customer, and use the Identify operation.. The Face API's PersonGroup and Identify operation allow you to create a group of VIP customers with minimal biometric data: each person is represented by a unique person ID derived from their face features, not the actual image. This complies with privacy regulations requiring minimal data storage. Option A is wrong because training a Custom Vision model requires storing and processing images, which retains excessive biometric data. Option B is wrong because Face ID is a temporary identifier that expires after 24 hours, making it unsuitable for long-term VIP identification. Option C is wrong because storing face images in Blob Storage retains the actual image, violating the requirement to minimize biometric data.

Answer

Train a Custom Vision model with images of VIP customers.

Answer

Use the Face API's Face ID returned upon detection and store it in a database.

Answer

Store the customer's face image in Azure Blob Storage and compare using the Face API.

Question 4

Refer to the exhibit. You are configuring an Azure AI Video Indexer job. The exhibit shows a JSON snippet of the job configuration. What will Video Indexer extract from the video?

Accepted Answer

OCR text, detected faces, and visual labels. The JSON snippet configures an Azure Video Indexer job with the 'features' array set to include 'ocr', 'faces', and 'labels'. This instructs Video Indexer to extract OCR text (from on-screen text), detected faces (via facial detection and recognition), and visual labels (objects, scenes, and actions). These are the only features specified, so the output will contain exactly those three types of insights.

Answer

Sentiment analysis of spoken content

Answer

Audio transcript and speaker diarization

Answer

Keyframe extraction only

Question 5

Refer to the exhibit. You are configuring the Azure AI Vision Analyze Image API. You need to ensure that the response includes a list of detected objects and brands. Which parameter is missing from the request?

Accepted Answer

Add 'model-version' with value '2023-10-01' to specify a stable version.. The Analyze Image API requires a stable model version to reliably return detected objects and brands. Without specifying 'model-version' with a value like '2023-10-01', the API may use a preview or default version that does not support brand detection or may return inconsistent results. Adding this parameter ensures the response includes the 'objects' and 'brands' visual features as expected.

Answer

Add 'smartcrops' to visualFeatures.

Answer

Add 'gender-neutral-caption' parameter set to true.

Answer

Remove 'Language' parameter because it is not supported.

Question 6

A retail company uses Azure AI Vision to analyze in-store video feeds. They want to detect when employees are not wearing required safety vests. The solution must send real-time alerts to store managers. Video feeds are captured from existing CCTV cameras. Which approach should you recommend?

Accepted Answer

Train a Custom Vision classification model and deploy it to the edge on CCTV cameras. To detect objects in real-time from live CCTV feeds, the model must run at the edge with low latency. Custom Vision supports object detection models that can be exported to run on edge devices such as cameras or IoT devices. This enables real-time inference without sending video to the cloud. Azure AI Video Indexer is designed for post-processing indexed videos, not live real-time detection. Azure AI Vision Image Analysis works on single images, not live video streams.

Answer

Use Azure AI Vision Image Analysis with a custom model to analyze each frame offline

Answer

Use Azure AI Video Indexer with a custom object detection model to detect vests in video frames

Answer

Use Azure Video Analyzer for Media to index videos and detect vests after recording

Question 7

A museum wants to create an interactive exhibit where visitors take a photo of a painting and the app provides information about the artist and historical period. The solution must use a pre-built AI service and be deployed on Azure. Which approach should you recommend?

Accepted Answer

Use Azure AI Vision Image Analysis to generate captions for the paintings. Azure AI Vision Image Analysis is a pre-built AI service that can analyze images and return tags, captions, and descriptions. For famous paintings, it may recognize the artist and period from pre-trained models. Option B requires custom training, which is not pre-built. Option C is for video, and D for documents. Therefore, A is the recommended approach.

Answer

Train a Custom Vision classifier with images of paintings labeled by artist and period

Answer

Use Azure AI Video Indexer to analyze the painting image

Answer

Use Azure AI Document Intelligence to extract text from the painting

Question 8

A logistics company uses Azure AI Vision to analyze images of packages on conveyor belts. They need to detect damaged packages and read tracking numbers. The solution must process high throughput (1000 images per minute) with low latency (<500ms per image). The images are captured by fixed cameras. Which approach should you recommend?

Accepted Answer

Use Azure AI Vision OCR Read API for tracking numbers and a separate Custom Vision model for damage detection. It combines Azure AI Vision OCR Read API (for text extraction) with a separate Custom Vision model (for damage detection). This leverages the specialized strengths of each service: the OCR Read API is optimized for text reading with high accuracy, while Custom Vision excels at object detection. The latency requirement (<500ms per image) can be met through parallel processing or edge deployment, and throughput can be achieved by scaling API calls. Option B is incorrect because Custom Vision does not include built-in OCR capability; it cannot read tracking numbers. Options A and C are less suitable: Document Intelligence is designed for structured documents, not real-time package analysis, and Video Indexer is for video streams, not still images.

Answer

Use Azure AI Document Intelligence to process package labels

Answer

Train a single Custom Vision model that detects damage and reads tracking numbers using OCR

Answer

Use Azure AI Video Indexer to analyze the video stream from cameras

Question 9

A retail company uses Azure Computer Vision to analyze customer traffic in stores. They deploy a custom object detection model to count customers and detect occupancy. After deployment, the model consistently underestimates the number of customers during peak hours. The company has retrained the model with more data but the issue persists. What is the most likely cause?

Accepted Answer

The training data does not adequately represent peak-hour scenarios.. The model consistently underestimates customer counts during peak hours, which indicates a distribution shift between the training data and the inference environment. Even after retraining with more data, the issue persists because the additional data likely still lacks sufficient representation of peak-hour scenarios (e.g., high density, occlusion, rapid movement). In Azure Custom Vision, object detection models learn from labeled examples; if the training set does not include diverse peak-hour images with varied lighting, crowd densities, and angles, the model will fail to generalize to those conditions.

Answer

The model is not being batch-processed for inference.

Answer

The model is overfitting to the training data.

Answer

The Computer Vision API version is outdated.

Question 10

A hospital uses Azure Custom Vision to classify X-ray images as normal or abnormal. The model achieves 98% accuracy on the test set. However, during deployment, the model misclassifies many abnormal cases as normal, causing missed diagnoses. The hospital has a class imbalance where abnormal cases are only 5% of the data. What should the data scientist do first to address this?

Accepted Answer

Use oversampling or class-weight techniques to balance the training.. The primary issue is class imbalance, where abnormal cases constitute only 5% of the data. Oversampling (e.g., SMOTE) or class-weight techniques adjust the training process to give more importance to the minority class, directly addressing the model's bias toward the majority class and reducing false negatives. This is a standard preprocessing step in Custom Vision and other ML frameworks before tuning hyperparameters or changing algorithms.

Answer

Increase the number of training epochs.

Answer

Add more normal X-ray images to the dataset.

Answer

Switch to a different object detection algorithm.