Deep Learning

Deep Learning is a type of machine learning method based on artificial neural networks. The algorithms used in Deep Learning are usually very complex and consist of several layers of neural networks.

Neural networks consist of a large number of neurons that are connected to each other and pass data from one layer to the next.

An artificial neural network learns by being trained on a large amount of sample data. During the training process, the weights and connections between neurons are adjusted to better fit the network to the sample data. After training, the network can be applied to new, unknown data to perform predictions or classifications.

The EYYESNet is a neural network for visual object recognition from images, covering the most important traffic-related classes (person, bicycle/motorcycle, car, truck/bus, train/tram) for automotive and railway solutions. It achieves the detection quality of a YOLO v3 object detector with a processing complexity of only 5%, enabling higher frame rates and multi-sensor setup with the same processing hardware.

EYYESNet can be used with traditional AI inference engines or the powerful EYYES Layer Processing Unit. It has been developed in a step-by-step training process over 5 years, acquiring more than 5 PB of real image data, plus more than 10 PB of synthetic data for different weather and environmental conditions.

Detection of:

• Pedestrians (adults, children, …)
• Cyclists
• Cars, trucks, trains
• Traffic signal detection
• Traffic sign recognition
• Lane detection