VECID – Number Plate Reading with Deep Learning

With the new VECID (= Vehicle Identification) tool the EyeVision software is able to read the number plates on cars or motor bikes, etc. with the state-of-the-art set of algorithms. The base of the VECID Deep-Learning-Algorithms is the TensorFlow open source framework and library for training our own models.

The pre-trained networks can be used to classify the image data.

The EyeVision provides Standard networks for applications such as:

• Number Plate Reading (NPR)
• OCR
• Make & Model (MaM )

The NPR (Number Plate Reader) combines the already known EyeVision OCR tool with the VECID algorithms. The OCR tool is only used for the separation of the characters. Afterwards the VECID NPR will work on the number plates captured e.g. at a car park entrance, toll roads or for law enforcement. It is much more powerful than the usual OCR tool as VECID Deep Learning is not susceptible to distortion, reflections or partial over- as well as underexposure, etc.

VECID will run on the following platforms:

• x86
• embedded ARM systems
• Windows
• Linux

In addition, due to the flexible hardware support by the EVOS (EyeVision Operating System) the Deep Learning System is instantly available for the small single core ARM processors up to the latest x86 platforms.

Market Intelligence News & Insights: 

How Will Apple’s $500 Billion Investment for US Manufacturing, Education, and AI Play Out?

On February 24^th, Apple announced a “commitment… to spend and invest more than $500 billion in the U.S. over the next four years”, with plans for new manufacturing plants, enhanced support for manufacturing partners via their Advanced Manufacturing Fund, and a myriad of other initiatives. As part of this plan, Apple projects 20,000 new hires, with roles focused on R&D, silicon engineering, software development, and AI/ML engineering.

Read More

Therefore with the new release EyeVision users will be able to train convolutional neural networks (CNN) on the base of Deep-Learning-Algorithms.
A CNN is composed of layers that filters (convolve) the inputs to get useful information. These convolutional layers have parameters (kernel) that are learned so that these filters are adjusted automatically to extract the most useful information for the task at hand without feature selection. CNN are better to work with images.

Another important point of using convolution as pattern match is that the position where the thing that is to be searched on the image is irrelevant.
There are additional AddOn commands for the EyeVision Deep Learning algorithms such as:

• OCR
• CoreFinder
• PeopleTrack
• ContainerFinder

The DL AddOns can be used with all EyeVision tools. ContainerFinder will be an AddOn command to easily find and identify container on, for example, a passing train. With the VECID tool the neural network learns from different images of containers to identify all container types to classify as "container“. Then the usual VECID OCR command can be applied, or alternatively the color commands, in case the container color is required.