Deep Learning from Adaptive Vision
10 years ago Adaptive Vision has redefined graphical programming for machine vision applications. Now it is time for another breakthrough – detecting defects with no programming at all.
By using Deep Learning technology the software tool is trained with Good and Bad samples, and then it automatically classifies input images as accepted or rejected. The new functionality speeds up the entire setup process of video systems. Standard training time is only 5 minutes on GPU, and typical applications require between 20 and 50 images for training.
The new software can work in two modes:
- Supervised mode – in this mode the user needs to carefully label pixels corresponding to defects on the training images. The tool then learns to distinguish good and bad features by looking for their key characteristics,
- Unsupervised mode – in this mode training is simpler. There is no direct definition of a defect – the tool is trained with Good samples and then looks for deviations of any kind.
Deep Learning works great with deformable objects, variable orientation and for example if customer provides vague specification with examples of Good and Bad parts.
Adaptive Vision was founded in 2007 as a new brand of Future Processing Sp. z o. o. company. Since then it have been providing machine vision software, libraries and development services.
The software is now available for sale. Please visit the product page for more information.
If you have any questions please don’t hesitate to contact AVICON sales team.