Highly skilled specialists to help your company with the design and implementation of AXIS ACAP applications. We will help you with education, system design, tools and implementation of edge applications for the Axis network cameras. No on-prem server is needed! We have long experience of all Axis ACAP SDKs that are used to write native applications that can be installed in the cameras.
The ACAP SDK from Axis gives you:
To build a product you need:
With our tools for automation and testing you can spend CPU cycles instead of valuable
head count. Our application package test tool that verifies the integrity of the .eap files is excellent for continuous integration (CI/CD). Combine it with our templates for GitHub Actions and you are up and running in no time!
Regardless if the outcome is a toolset, a software deployment unit or a development process, it will be custom tailored and applicable to solve your specific needs. With many years of history within the computer vision and video analytics fields for Axis' cameras, we have made solutions for domains ranging from retail analytics and industry 4.0 to IoT and building automation. By leveraging our experience and know-how together with our in-house development tools and libraries we help you deliver fast results with major impact for your business problems.
Are you getting into a new domain within ACAP development, Machine Learning, Computer Vision, MLOps or camera analytics? Our courses, videos and selection guides helps your developers get a flying start and orient themselves in the complex environment. With our tools, getting productive has never been easier!
Are you already in the game and experiencing growing pains or just looking for ways to make your systems scale better? We can help your organization with design discussions, problem solving and system wide recommendations.
Do you have a project idea which needs a Proof of Concept (PoC) or demonstrator? Are you committed to deliver a project with a tight dead-line? As a hired gun on small but critical projects we can do application bring-up, algorithm implementation, benchmarking and optimization or prototyping.
You can use us as your primary resource for ACAP development projects or you can request our services a few hours every week or month to deal with the most critical tasks. With our assistance your development team will never be stuck again. We are not requesting you to commit to paying for a certain amount of hours, instead our contract with you simply means that we are ready to help you whenever you need our experience.
Our clients are using our ACAP applications for everything ranging from simple rule based event filtering to advanced video analytics on the device. Our core focus is Machine Learning (ML), Deep Learning (DL) and Computer Vision (CV) for the Axis network cameras using the modern tool stack for Machine Learning Operations (MLOps) and Amazon Web Service (AWS) together with the Axis Camera Application Platform (ACAP).
Rule-based triggers and health monitoring
Simple applications on your camera that monitors its status or reads its event streams. The applications are preconfigured to detect specific conditions such as combinations or sequences of events. When a condition is met the application produces a notification as a new event, HTTP push message or any other means of communication that can be received by a VMS or server.
Applications installed on the camera that can upload image or video data to a cloud provider, where video analytics are performed. These applications can also fetch configuration data from the cloud to customize the camera's and the application's behavior. The edge application can host a web-based user interface that allows any operator to configure or troubleshoot the application.
Edge ML applications use deep learning models or machine learning algorithms directly on the camera without requiring any additional hardware at the site. This can be achieved by using pre-trained models from a model zoo or by training custom models using TensorFlow. Some of the cameras has special hardware for ML (DLPU), while the other cameras can use CPU-based models. Edge ML can generate aggregated statistics or real-time event triggers based on the model output.
Data storage, model training and model compilation is done in the AWS cloud. Once the models are trained and evaluated they are deployed to the edge camera systems by integrating them in an ACAP application that is installed in an Axis camera with a Deep Learning Processing unit (DLPU), e.g. the ARTPEC-8 platform. Collection of training data is performed as a separate task on all or selected devices.
Both data storage, training and model inference is performed in the AWS cloud environment. Data from connected Axis cameras is streaming or batch-uploading to a cloud gateway where the model prediction is performed. Some of the data is directed to the data lake to use as training data for the next iterations of training jobs.
We are working together with your own development team to build in-house competence in your organization.
Recommend libraries, integrations and system designs
Create project breakdown and perform time and cost estimates
Create initial ACAP application structure and customized tools
Empower your developers with knowledge about the Axis ecosystem
Retained advisory or consultancy
Hands on workshops mixed with conceptual understanding will help your developers regardless if they are new to video analytics or seasoned computer vision experts. Our company trainings comes in different flavours and can be tailored to your needs.
Focus on the most value generating proof-of-concepts by giving your developers an extensive introduction to ML and the concepts around it. The course can be given either at a decision maker level or at a developer level.
Growing pain, slow lead times or failure to scale PoC models can be combatted with the mapping and introduction of MLOps best practices. This training is best tailored to your specific organisation and can be given either at a decision maker level or at a developer level.
Getting started with ACAP development or migrating to a new SDK can seem daunting, but there is a clear structure to how the environments are structured. Depending on your use case, one SDK will be better than another. With hundreds of AXIS camera models to choose from you need to understand where they differ and what properties they share.
Don't miss any upcoming insights, register on the mailing list for webinars and white papers.
© 2021 FixedIT Consulting AB
VAT: SE 559296849801
Org. Number: 559296-8498
Phone: +46 76 80 156 12
Email: Daniel.Falk.1@FixedIT.ai
