Contacts:
algos@winkam.com



About WINKAM

OUR MISSION. As an R&D company, we are committed to doing research, experiments, and prototypes and delivering industry state-of-the-art-accurate detection algorithms for our clients’ products in digital sports, medicine, mobile games, and autonomous driving. We help companies to develop accurate algorithms for their products to extract meaningful data from raw sensor readings. WINKAM’s Technology is a special blend of math, a love of sensors, and a drive to provide the highest possible accuracy of real-time recognition.

TECHNOLOGY LAB. Our lab focuses on proprietary research and development for real-time pattern recognition using input from diverse sensors (e.g., MEMS accelerometer, gyroscope, RGB video camera, magnetometer, etc). We do all kinds of Machine Learning and Neural Networks as well as handcrafted algorithms. To create a solution, we apply our knowledge in math, algorithms, sensors’ physical limits and context specific to use case. We then enhance the algos with machine learning techniques. Finally, we pack all of it into software libraries for convenient integration into products.

QUALITY ASSURANCE. Testing, checking, and quality control are essential processes at Winkam. We genuinely want products we are part of to satisfy and bring value to end-users. To achieve high precision and accuracy, we have our own data collection and labeling department. This helps us to prepare testing and training data sets carefully and representatively to reality. So, our algorithms work in real-life conditions, not just in theory. As a quality guarantee, we provide support services to enable long-term fault-free performance.

POLICY: CONFIDENTIALITY AND SECURITY. WINKAM is an IP holding company. We protect our intellectual property and respect the intellectual property rights of others. We have a responsibility to protect and not to leak Winkam confidential information or the confidential information of our customers and business partners.

If you are looking for a partner to take care of algorithm development for your product or you just need advice ㄧ contact us at algos@winkam.com

Recent algorithm applications

WINKAM in Road Safety

Current autonomous-driving systems almost ignoring cyclists and creating dangerous situations while in proximity with cyclists has raised the alarm in the industry to focus more on cyclist detection. To address this, we conducted several scientific experiments on potential cyclist detection improvements for autonomous vehicles and developed the Cyclist Detection Classifier — CyDet.

CyDet uses an input of mono video camera, reaches state-of-the-art Average Precision of cyclist detection on KITTI data set, and has a real-time processing rate of 60 FPS on a 35W Intel Core i7 which is regular laptop CPU.

Demo at IEEE IV2018, Intelligent Vehicle Symposium 2018.
See pipeline video See scientific research overview

In collaboration with





WINKAM in Sports

Hm, how can I make my jumpshot more consistent? What should I focus on in my training to really be better in defense? Do I accelerate enough with my dribbling burst? To answer all these questions and get more actionable insights for coaches and players we developed an algorithm library for Basketball Movement Patterns and Metrics recognition — the MRBC.

MRBC identifies movements invisible to the naked eye in real-time and digitises them into metrics to generate insights. Our multi-sensor algorithms use input from three 3-axis accelerometers and three 3-axis gyroscopes attached to players. The MRBC recognizes 11 movement patterns and metrics such as dunk power, jumpshot release time, dribbling burst acceleration and defense reaction time.

Live demo at CES 2017

Looking for sports algos? Contact us at algos@winkam.com





WINKAM in Mobile Games

Gamers need more buttons! Fingers tend to take up too much screen space. The thumbs cover about 35% of the screen reducing visible gameplay area. This is a critical issue for dynamic games. We have a solution for a more flexible UX.

Winkam can put buttons outside the screen allowing for the Gamepad-like experience. Enhance your mobile gameplay. Integrate SDK to add full-size gamepad controls to a smartphone.

Learn more




WINKAM in Signal Transformation

We know how painful it is to work with corrupted UWB XYZ signal and how signal random jumps and distortions destroy precision of good algorithms.

That’s why we developed our proprietary methods of signal random-jump filtering which improves UWB signal quality.

Would you like to find out more? Contact us at algos@winkam.com