What is Azure Percept? Understanding Its Features

What is Azure percept? How does it work and who will be using It? Learn its basic components

Before the introduction of Azure Percept, most of the edge AI solutions and models had to be built from the ground up and required users to have extensive experience in data science and cloud computing. As a result, building and deploying an AI model took a lot of time and was a challenging task.

With Azure Percept, Microsoft solved these difficulties by providing an easy-to-use AI model platform bundled with the necessary hardware and software. Read more below to learn about Azure Percept and its unique features.

Azure Percept

  • Azure Percept is a bundle of software, services, and hardware that is specifically developed to enable business transformation using modern tools such as AI and IoT at the edge. Azure Percept aims to solve integration difficulties and challenges of edge AI at scale and cover the full stack from silicon to services.

Azure Percept allows efficient development and building of custom AI models, management and setup of IoT devices, and seamless integration with Azure cloud services. All this is to enable at-scale deployment and AI application prototyping.

  • The learning curve is significantly lowered for new users as Azure Percept is highly compatible with existing IoT architectures. In addition, it offers low-code development, thus allowing users with limited data science knowledge to work on it with complete confidence.

Components of Azure Percept

Azure Percept has three core components that it provides support via its platform. They are on-device prototyping (development kit), guidance for best practices, and cross-device solution management and workflow (portal).

  • Azure Percept Development Kit offers diverse AI prototyping scenarios utilizing computer speech or vision through varied and flexible hardware options. Security solutions and hardware acceleration are provided as inbuilt features. In addition, it offers efficient and real-time insights for AI models that can run on the device without requiring a cloud connection.
  • Azure Percept Studio is a management and launch portal for all edge AI models and solutions. Percept studio offers pre-developed models and allows users to create their models. By utilizing its low-code development features, any user can quickly and efficiently develop new models without needing expertise in data science or cloud computing.
  • AI Hardware Reference Design and Certification Programs offer best practices guidances, data sheets and device specifications, assistance for device deployments, and help in firmware updates.

Features of Azure Percept

  • One core feature of Azure Percept is that it is an edge-edge AI platform that offers built-in hardware acceleration integrated with Azure AI and IoT services. It allows users to initiate their projects using pre-developed AI models such as shelf analytics, vehicle analytics, object detection, etc. Users can further customize these models or create their own effortlessly.

Azure Percept Studio offers an unmatched experience for creating edge AI solutions. Its platform offers an integrated collection of Azure services and tools that can be used for device provisioning, AI and machine learning deployments, and AI lifecycle management at the edge.

  • Azure Percept offers the best security features and solutions designed to protect your sensitive and highly valuable assets. Its hardware accelerators are designed with a hardware root of trust that interacts with device TPM and attestation services to prove device identity. In addition, the zero-trust security module enables absolute security for AI models, data in transit, and edge AI solutions.

Azure Percept Sample AI Models

  • what is azure perceptAzure Percept Studio currently offers AI sample models for vehicle detection, people detection, object on-shelf detection, and general object detection. There is no requirement for prior knowledge of advanced data science and cloud computing. A user can simply deploy the model to Azure Percept DK and can watch the model work in real-time using the devkit’s video stream.

An open-source people counting solution is available for all users that work with AI. It helps users count the number of people within a specified zone and time range. An on-premise edge device is used that egresses data to Azure Data Lake, which is shown on the user interface.

  • Custom vision and speech solutions can also be developed using Azure Percept’s pre-developed AI models. These can be trained using user-defined images. Azure Cognitive Services is used for training and evaluating the models. Vision and Speech AI models have applications in healthcare, tourism, inventory, and automotive.

If you are interested in availing of the services of Azure Percept, then get in touch with Henson Group today. We are one of the best-managed service providers (MSP) for Microsoft Azure and have a robust global presence. Our dedicated team will help you with all your Azure-related needs.