Identify capabilities of the Computer Vision service

Computer Vision is an AI service within Microsoft Azure that provides developers with advanced algorithms that are designed to process images and return information based on the visual content. These capabilities can be broadly categorized into the following areas:

Image Analysis:

The Computer Vision service can analyze visual content in different ways, providing insights such as:

• Tagging: Automatically identifies and tags visual features in an image. For example, an image of a cityscape might result in tags such as “building,” “skyline,” and “city.”
• Description: Generates a human-readable phrase that summarizes an image. For example, the service might describe the aforementioned cityscape as “a panoramic view of a city skyline at sunset.”
• Category: Classifies images into thousands of categories like “outdoor” or “people.”
• Object Detection: Recognizes and locates objects within an image. For instance, it might identify and locate “cars” and “pedestrians” in a street scene.
• Branding: Detects and identifies commercial brands based on logos, products, or storefronts.
• Faces: Detects human faces and returns coordinates, gender, age, and emotion. However, it doesn’t include facial recognition capabilities.
• Color: Identifies dominant background and foreground colors, including the accent color.
• Adult/racy content: Evaluates visual content for adult or racy material and can provide a safety rating.
• Image Types: Detects image types such as Clip Art, Line Drawing, or a Photo.

Optical Character Recognition (OCR):

This allows for the detection and extraction of text within images. OCR can be used for:

• Text Extraction: Reads text from images, such as photographs of documents, and provides a text stream.
• Handwriting Recognition: Analyzes handwriting in images and translates it into machine-encoded text.
• Multi-language Support: Can recognize and extract text in various languages, expanding the usability of the Computer Vision service globally.

Spatial Analysis:

Utilizes spatial analysis to understand the movement of people within a space, useful for a variety of scenarios:

• People Counting: Estimates the number of people in a particular area.
• Movement Patterns: Analyzes how people move through a space, which can be utilized for optimizing store layouts or evaluating the flow of foot traffic.
• Social Distancing Analysis: Assesses the compliance with social distancing guidelines in a physical setup.

Customization and Training:

With the Custom Vision service, a subset of Azure’s Computer Vision, users can train a model to recognize specific content in imagery relevant to their business needs. This includes:

• Custom Image Classification: Used to categorize images according to specific criteria defined by the user.
• Custom Object Detection: Personalizes the service to identify and locate unique objects.

Domain-specific Analysis:

Offers models tuned for specific domains:

• Celebrities and Landmarks: Detects and identifies celebrities and landmarks from a database of noteworthy individuals and natural or man-made structures.
• Retail: Tailored recognition abilities geared towards retail applications, such as inventory management based on image content.

Let’s consider a comparative overview of Key features of the Computer Vision Service:

These capabilities enable developers to build sophisticated computer vision scenarios, ranging from moderating content and processing forms, to recognizing objects in the physical world. Applications built on Computer Vision can be used across a variety of industries, including retail, manufacturing, health care, and public safety.