We think that you are aware of the existence of a useful Intel product called the Open Visual Inference & Neural Network Optimization (OpenVINO) toolkit - a set of libraries, optimization tools, and information resources for developing software that uses machine vision and Deep Learning. Also, you are probably aware that the best way to learn a tool is to try to do something from scratch with it. If both theses do not cause you to object, then you are mentally prepared to take part in the OpenVINO hackathon, which Intel holds in Nizhny Novgorod from November 30 to December 1 .
And Nizhny Novgorod, by the way, is less than 4 hours "Swallow" from Moscow. This is another argument in favor.
Everyone is invited to participate in the hackathon, having at least some programming experience in C or Python. You can register immediately with the whole team, or you can singly - everyone will be interested.
Any ideas are accepted for development. The task of the teams of participants is to propose the use of computer vision algorithms based on neural networks to solve one or more applied problems. In addition to the idea, it is required to demonstrate the prototype of the solution or its parts using the Intel OpenVINO toolkit software product, as well as to evaluate the complexity of implementation and deployment.
Examples of directions and topics
Safety
Retail and Entertainment
Industrial
- Detection of anomalies in human behavior: aggressive movements, wandering (the duration of a person's presence on the stage is suspicious), falls (medical help is needed).
- Roadside assistance: monitoring the driver’s condition, analyzing the traffic situation, predicting emergency situations to prevent them, detecting and recognizing license plates.
Retail and Entertainment
- Computing photo. Using convolutional neural networks to improve image quality / post-processing. Integration of deep learning solutions with web services (chat bots, Web GUI).
- Advisory and adaptive systems based on predictions of gender, age, emotions and other user attributes.
- Recognition of visitors, tracking indoors, analysis of time spent and areas of visit.
- Human posture assessment: sports trainer, 2D and 3D skeletal animation, gesture control.
Industrial
- Smart factories and enterprises: industrial safety control (abandoned tools that are forbidden to visit the territory), process automation, anomaly detection.
- Deep learning for home: security systems, auxiliary devices
- Agriculture: detection of pests, plant diseases.
Each team will be provided with a Raspberry Pi 3 board and Intel Neural Compute Stick 2 hardware accelerator for the duration of the hackathon. On work laptops, it is recommended to install the Intel OpenVINO toolkit in advance and check its operation.
The winners of the hackathon will receive cash prizes: for 1st place - 100,000 rubles, for 2nd place - 50,000, for 3rd place - 30,000 rubles.
So, we meet in the morning of November 30 in Nizhny Novgorod at the address Pochainskaya street, house 17 building 1. Come and come to visit!