Stepan Novikov

I have 14+ years of experience in development of software projects, with more than 7 years’ experience with enterprise level system architecture acting as an architect/team lead. I have solid understanding and knowledge of object-oriented approach, software engineering principles and concepts. I have experience in leading projects, communication with clients, risk mitigation. I have performed analysis of customer environment, design strategies and technical architecture and project implementation. My primary focus were: the banking software (bank systems, credit card processing systems, etc); performance optimization in enterprise level systems; machine learning data science research for recognition/prediction

Pattern recognition & computer vision

Pattern recognition & computer visionHow to recognize a person on photo with high accuracy in real-time? How to detect defect automatically? How to create self-learning predictive analysis system?You will find answers to all of these questions with practical examples of realization in .Net and accuracy up to 99%Technologies: Artificial neural networks and Genetic algorithms, Scanning windows and Yolo. 

Artificial neural networks: areas of use and learning methods on base of perceptron like network structure

The key idea of a pattern recognition approach is to develop a computer system (CS) ability to adapt to the changes in environment to perform the tasks. You can do this by creating a basic set of samples of objects and actions in the world and a set of possible responses to them. Based on these sets of CS can build your own strategy or adapt of the existing one. Then CS can use it to perform tasks under the circumstances. Agenda: 1. Universal pattern recognition approach presentation 2. Practical experience for automation of detection of defects 3. First steps with big datasets (40k+ samples). Digit recognizer.