PRODUCTS AND TECHNOLOGY
We follow one fundamental approach
Model-free approach: by using supervised or unsupervised neural networks to identify the process fault/no-fault state. Our approach based on CNN do not needs any modelling assumptions.
About our Products
Image classification service by Convolutional Neural Network
Data Set Creation:
Automatic data- set creation by multi-pose image acquisition and image labeling
Neural Network Training:
Training on NVIDIA® hardware GPU
Customized Hardware Integration for AI:
Integration of our Artificial Intelligence engine in custom hardware in order to meet the desired speed of execution
Hand-on tablet for image recognition
Windows 10 tablet with embedded AI algorithm for image recognition:
Our standard solution is based on Windows Surface Pro 4, 7, X tablet specific for hand-on mobile applications
VARTM (Vacuum-Assisted Resin Transfer Molding) MONITORING BY SURROGATE MODEL BASED ON CNN
AIfScope: AI front Scope, tool to track the resin front based on binary sensor (resin presence), click here.
About out Tools
We work with the best technologies available on the market.
The image is captured from four different points to have greater certainty of the results, using a tablet surface with a rear camera.
Digital-twin development
Solidworks Flow Simulation: a computational fluid dynamics analysis solution to simulate the general flow through the finite volume method (FVM), calculate the performance and optimize the results
Artificial Intelligence application
- Matlab Deep Learning Toolbox: to design, improve and analyse convolutional neural networks with deep learning algorithms
- Parallel Computing Toolbox: to solve high data content problems using multicore processors, GPU and cluster
- Matlab Compiler: to make standalone and web app sharing Matlab programs