Week 4: Modern Computer Vision Applications Flashcards
What are the major components involved in handwritten digit recognition?
Handwritten digit recognition involves preprocessing steps like normalization and segmentation, feature extraction to capture essential patterns, and classification using algorithms like neural networks or support vector machines.
How is the performance of handwritten digit recognition evaluated?
The performance of handwritten digit recognition is typically evaluated using metrics like accuracy, precision, recall, and F1-score. Accuracy measures the ratio of correctly recognized digits to the total, while precision and recall assess the model’s ability to correctly identify positive instances and capture all positive instances, respectively.
How can you utilize Microsoft Azure Computer Vision services?
Microsoft Azure offers Computer Vision services that enable image recognition tasks such as object detection, image analysis, text extraction, and content moderation.
What capabilities does Microsoft Azure offer for object detection?
Microsoft Azure provides tools and services for training object detection models, facilitating tasks like identifying and localizing objects within images or videos.
How does the Microsoft Video Analyzer support object detection tasks?
The Microsoft Video Analyzer enables users to perform object detection on videos, allowing the identification and tracking of objects across video frames.
What are the three main categories of image features?
Colour moments, shape features, and texture features.
What is the purpose of feature extraction in image classification?
To extract relevant information from images that can be used to distinguish between different classes.
What are some examples of colour moments?
Mean, variance, skewness, and kurtosis of the colour channels.
What are some common shape features?
Aspect ratio, perimeter, compactness, and roundness.
What are some types of texture features?
Statistical features (e.g., mean, variance), local binary patterns, and Gabor filters.
What is K-nearest neighbours (KNN)?
An instance-based classification model that predicts the class of a new example based on the classes of its nearest neighbours in the training data.
What are the advantages and disadvantages of KNN?
Advantages: Simple to implement, few parameters to tune.
Disadvantages: High computational cost for classification, sensitive to irrelevant features.
What are the four main performance metrics for image classification?
Accuracy, precision, recall, and F1 score.
What is mean average precision (mAP)?
A metric used to evaluate object detection models, taking into account the confidence scores of the detections.
What are the six principles of Responsible AI?
Fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability.
What is Microsoft Azure Machine Learning?
A cloud platform for building and deploying machine learning models.
What are some of the computer vision services offered by Microsoft Azure?
Computer Vision (object detection, image description), Custom Vision (training custom models), Face (facial analysis), Form Recognizer (extracting information from forms), and Computer Vision (reading text).
