Microsoft Azure AI Fundamentals: Computer Vision PDF

Microsoft Azure AI Fundamentals: Computer Vision © Copyright Microsoft Corporation. All rights reserved. Agenda Computer vision concepts Computer vision capabilities in Azure © Copyright Microsoft Corporation. All rights reserved. Learning Objectives After completing this module, yo...

Microsoft Azure AI Fundamentals: Computer Vision © Copyright Microsoft Corporation. All rights reserved. Agenda Computer vision concepts Computer vision capabilities in Azure © Copyright Microsoft Corporation. All rights reserved. Learning Objectives After completing this module, you will be able to: 1 Understand the capabilities of Azure AI Vision. 2 Identify the different services included in Azure AI Vision. 3 Describe the Face detection service. © Copyright Microsoft Corporation. All rights reserved. Computer Vision Concepts © Copyright Microsoft Corporation. All rights reserved. Images and image processing An image is an array of pixel values Filters are applied to change images 0 0 0 0 0 0 0 0 -1 0 -1 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 -1 0 8 0 -1 0 0 0 0 0 0 255 255 255 0 0 0 -1 0 -1 255 -1 255 255 0 0 0 0 255 255 255 0 0 0 0 255 0 255 0 0 0 0 255 255 255 0 0 0 0 255 255 255 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 © Copyright Microsoft Corporation. All rights reserved.  Convolutional Neural Networks 1. Labeled mages are used to train the model 2. Filter layers extract feature maps from each image 3. The feature maps are flattened 4. The feature values are fed into a fully connected neural network 5. The output layer produces a probability value for each possible class label During training, the filter kernels start with random weights. These weights are iteratively adjusted to improve the accuracy of the predictions based on the known labels. The trained model uses learned weights to extract features from new images and predict their class. © Copyright Microsoft Corporation. All rights reserved. Multi-modal models A newer approach to modeling involves combining language and vision models that encode image and text data The model encapsulates semantic relationships between features extracted from the images and text extracted from related captions. A multi-modal model can be used as a foundation model for more specialized adaptive models. © Copyright Microsoft Corporation. All rights reserved. Computer vision services in Azure Vision Face Image Analysis: Face detection Image tagging, captions, Face recognition model customization, and more. Optical Character Recognition (OCR) Spatial analysis © Copyright Microsoft Corporation. All rights reserved. Computer Vision Capabilities in Azure © Copyright Microsoft Corporation. All rights reserved. Image analysis 4.0 with the AI Vision Service Capabilities include: Model customization person Read text from images person male (34) Detect people in images Generate image captions jeans Detect objects footwear Tag visual features Smart crop Caption: A group of people walking on a sidewalk Tags: Building, jeans, street, outdoor, jacket, city, person © Copyright Microsoft Corporation. All rights reserved. Exercise: Analyze images in Vision Studio 1. Use the hosted environment and Azure credentials provided for this exercise. In this exercise, you will use the 2. The instructions are also available on Learn: Azure AI Vision https://aka.ms/ai900-image-analysis service to analyze images. © Copyright Microsoft Corporation. All rights reserved. © Copyright Microsoft Corporation. All rights reserved. Detecting faces with the Face Service Everyone can use the Face service to detect:  Blur  Exposure  Glasses  Head pose  Noise  Occlusion Only Managed Microsoft customers can access facial recognition capabilities:  Similarity matching  Identity verification © Copyright Microsoft Corporation. All rights reserved. Demo: Detect faces in Vision Studio 1. Follow along on the exercise page at: https://aka.ms/ai900-face In this demo, you will take a look at the face detection capabilities of the Azure AI Face service. © Copyright Microsoft Corporation. All rights reserved. © Copyright Microsoft Corporation. All rights reserved. Reading text with Optical Character Recognition (OCR)  Detect the location and characters of printed and handwritten text  Options for quick text extraction from images, or asynchronous analysis of larger scanned documents Shopping List Non-fat milk Bread Eggs © Copyright Microsoft Corporation. All rights reserved. Demo: Read text in Vision Studio 1. Follow along on the exercise page at: https://aka.ms/ai900-ocr In this demo, you’ll see Azure AI Vision’s optical character recognition capabilities in action. © Copyright Microsoft Corporation. All rights reserved. © Copyright Microsoft Corporation. All rights reserved. Knowledge check You want to use the Face detection service to identify faces in images. What can be identified using the Face 1 detection service? ⃣ Faces that cannot be seen because the person has turned their back. ⃣ Partially obscured faces. ⃣ Faces that are obscured by another object. You want to use the AI Vision and AI Language service. You also want developers to require only one key and 2 endpoint to access all your services. What kind of resource should you create in Azure? ⃣ Azure AI service ⃣ Language ⃣ Vision 3 Which services are part of Azure AI Vision? ⃣ Face detection and speech recognition ⃣ Optical Character Recognition and face detection ⃣ Document Intelligence and speech recognition © Copyright Microsoft Corporation. All rights reserved. References Read more about: Fundamentals of Computer Vision Fundamentals of Facial Recognition Fundamentals of Optical Character Recognition Through the content on Learn: Microsoft Azure AI Fundamentals: Computer Vision - Training | Microsoft Learn © Copyright Microsoft Corporation. All rights reserved. © Copyright Microsoft Corporation. All rights reserved.