UNIQUE PEOPLE AND COMPETENCES
We are a deep-tech company which offers cutting-edge machine learning expertise and solutions that are tailored to your specific application and to your current software and hardware environment.
Our machine learning solutions are always based on state-of-the art research and our system architectures are future proof and ensures stability, uptime as well as software security.
Our system solutions can be integrated in your current machine and manufacturing process but in case you would like to upgrade your factory with new hardware with vision or other sensor solutions or robot cells we collaborate with our system integrator partners to be able to offer turnkey solutions.
PACKAGED AI SERVICE:
We focus mainly on end users who would like to make sense of sensor data or image data of different kinds. With the growing field of IoT and transactional data we see that the use of AI and particular machine and deep learning is needed to be able to work with so large amounts of data.
Our team’s broad expertise covering both industrial manufacturing, data science and machine learning enable us to understand the demands for improved quality, productivitiy and process efficiency.
INFRASTRUCTURE:
We deploy machine learning solutions in your current production environment including architecture and integration to your IT-infrastructure.
We support you in updating and changing the models in use to ensure the solution is always improved and state-of-the art.
We offer data integration framework and storage for customers that would like to outsource this to us. We can also offer service agreements to you that would like to secure maximum uptime of your services.
KEY COMPETENSES:
OBJECT DETECTION, LOCALIZATION AND CLASSIFICATION
Object detection has a wide range of applications including anomaly detection, object counting, object tracking, video surveillance and pedestrian and vehicle detection for self-driving cars. Recent scientific progress in deep learning has resulted in a number of powerful machine learning algorithms for object detection that opens up the possibilities for new and interesting applications.
CROWD COUNTING
It can be about counting bees, but the same technique can also be used for other objects such as people, vehicles, trees, crops or products in a manufacturing process.
OBJECT DETECTION
Which are the infected and sick bees? This is one example of object detection. The same technique can be applied to detect patterns and anomalies in a wide range of applications, from transportation to security and manufacturing.
DEEP LEARNING IMAGE CLASSIFICATION
Image classification for faster and more accurate decision making. In this case replace or repair. The same technique can be applied to quality inspection in different manufacturing environments or to analysis of medical images. Amongst others.
ANOMALY DETECTION USING CAMERAS OR OTHER SENSORS
Anomaly detection is the task of finding patterns in data that do not conform to a notion of normal. This is relevant to several use cases such as manufacturing defect detection, medical diagnosis, fraud detection and more. Flagging unual cases when they occur can save business time and costs and increase customer satisfaction.
Example of anomaly detection system identifying four different faults on medical capsules:
Crack
Poke
Scratch
Squeeze
The top row shows the original input image; middle row shows ground truth, and bottom row the anomaly detection results: Location and severity, where red color means more severe faults.
Anomaly detection is used for quality inspections in many industrial processes. Read more about what our Vision Platform can do…
Image from THE MVTEC ANOMALY DETECTION DATASET
REAL TIME OBJECT TRACKING AND POSITIONING
Object tracking requires that the position of a moving object is updated and maintained in various scenes. It has a wide range of applications including process optimization where you are able to identify and correct bottlenecks and overrynning processes in real-time and to identify patterns to optimize processes, layouts, task sequencing, asset utilization and materials routing.
Another area of applications is safety and compliance; by using realtime object tracking you will be able to detect, locate and count products, devices and people real-time and by that identify and correct unwanted behaviours, prevent accidents, hazards, thefts and lost objects.
RECOMMENDER SYSTEMS AND REINFORCEMENT LEARNING
Which ads and products should be promoted to which customers and when?
By using AI models for recommender systems and reinforcement learning you will be able cluster customers and optimise your marketing activities.
The solution will learn and improve autonomously over time.
PREDICTIVE FORECASTING USING
MACHINE LEARNING TIME SERIES ANALYSIS
SALES, VISITORS PREDICTION
How many customer’s will come next week, next month, next quarter? How much staff do you need at a certain hour o the day, week, month?
By applying machine learning models to historic data and combining this with other data inputs like weather, holidays, marketing activities etc you will be able to predict future sales and vistors with better accuracy.
PRODUCTION PLANNING
By understanding and see patterns in massive amounts of data - both internal and external - you will be able to improve your forecasts and optimize your production planning and thereby your internal productivity.
These techniques can be used in most industry segments; from energy to manufacturing and process industries.