Microsoft AI School - Creating Magic with Vision Computing

Introduction

Computer Vision is a part of Artificial Intelligence that enables AI systems to visualize the world as a human by identifying and analyzing the content in images and videos. Azure offers various Cognitive Services and tools that allow developers to add AI capabilities to their applications without requiring in-depth knowledge of AI algorithms. These services can be used to extract rich information from images and videos to address common computer vision problems. To help you master these concepts and create computer vision solutions in the Azure platform, MS AI school features a perfect roadmap in the form of a learning path which you can pursue by following this link: "Create computer vision solutions with Azure Cognitive Services."  

Prerequisites

To pursue this learning path, you must have : 

• Basic understanding of Azure and Azure Portal.
• Proficiency in C# or Python programming language. If you do not have any prior experience in programming, please feel free to complete the Take your first steps with C# or Take your first steps with Python learning path first.

This learning path has four modules. Read ahead to get a quick peek at each module of this learning path. 

Module 1: Analyze Images 

Analyzing Images involves interpreting the visual information from an image. The image is analyzed in detail to identify the underlying objects in it. Thus, in the first module, you will learn how to use Azure's cognitive service named Computer Vision to accomplish this task with the help of pre-trained models. 

After completing this module, you'll be able to:

• Provision a Computer Vision resource to identify objects, descriptions, known brands, metadata, etc.
• Analyze an Image using REST or equivalent method in SDK.
• Create a thumbnail with smaller dimensions.

Here's an overview of the units covered in this module: 

Module 2: Analyze Video 

Video Analysis is a part of Artificial Intelligence that deals with interpreting the content embedded within the videos. Some use cases of Video analysis are detecting the location shown in the video and displaying relevant advertisements, identifying inappropriate content in the video and performing specific actions, etc. Microsoft Azure offers a cognitive service known as Video Analyzer for Media to analyze and extract insights from video content, including scene segmentation, face identification, text recognition, etc. 

Thus, after completing the second module, you'll learn to use the Video Analyzer for Media service to: 

• Extract information from the videos like detecting a face or a person in the video, reading optical characters, etc.
• Video Analyzer for Media service to create Custom models to identify a new person or language, brands, etc.
• Services like Video Analyzer for Media widgets and APIs.

Here's an overview of the units covered in this module: 

Module 3: Classify Images

Image Classification is a computer vision task where the software identifies the subject in the image, analyzes it, and categorizes it to either an already existing model or a custom-created model. Some real-world use cases are - identifying the number plates of vehicles, identifying the items in the shopping basket, self-checkout desks, etc. 

Microsoft Azure offers a cognitive service named Custom Vision service to analyze and categorize the images and after completing this module, you'll be able to:

• Build a custom model for image classification.
• Train the custom model to classify the images

Here's an overview of the units covered in this module: 

Module 4: Detect Objects in Images

Object Detection is a popularly known challenge in AI that involves identifying the location of an object in an image and categorizing it to an existing model or a custom-created model. Some use cases of Object Detection are - detecting cars to analyze the traffic, detecting people in Surveillance cameras, detecting animals in agricultural lands, etc. Azure's Custom Vision service allows you to locate and count the occurrences of objects in an image. Thus, after completing this module, you'll be able to: 

• Build a custom model for object detection
• Train the custom model to detect the presence and location of an object
• Use different options for labeling images.

Here's an overview of the units covered in this module: 

Module 5: Detect, Analyze, and Recognize Faces

An important Artificial Intelligence competency is the capacity for programs to detect human faces, assess facial traits and emotions, and identify individuals. You can use Azure's Computer Vision cognitive service that enables you to design solutions to detect a face. At the same time, Face service can be used to recognize facial features and to identify a face from multiple images of the same person.  

Hence, after completing the final module, you will be able to use the services as mentioned earlier to:  

• Detect the presence of a Face
• Recognize the Facial Features of a Face like an eye, hair, nose, etc.
• Train the Facial recognition model to Identify the Face.

Here's an overview of the units covered in this module: 

Conclusion

This learning path will guide you to understand the common problems faced in the field of Computer Vision and implement the various services offered by Microsoft's Azure Cognitive Services to come up with solutions to these problems even without having an in-depth understanding of complex AI and ML algorithms.