Improving Golf Performance Using Computer Vision
Technology has been advancing incredibly in all fields over the past five years be it in sports, media, healthcare, or any other domain. Interestingly, everything is becoming dependent on data.
Analytics in sports tech has numerous applications that can support professional training, in-game decision-making, advanced sports statistics, injury prevention, talent recognition, etc. Also, visual learning has always captured more attention and information and has more retention power. Therefore, we at Algoscale leveraged the power of Deep Learning in Video Analytics to explore how we can improve sports training. Deep Learning has found its way to a number of applications over the last few years.
An important reason for the same could be the advent of new infrastructure and libraries offered by tech giants like Google, Amazon, and others.
Algoscale worked with an upcoming startup to help them build their product that lets a Golf beginner compare their shots to that of a professional player. The solution used deep learning to learn the shot style on a video shot by the player and benchmark it against a host of professional videos. This helped the trainee become a better player with time.
The solution involved the use of the Fast R-CNN algorithm for training the model, and computer vision techniques and advanced mathematics to compare the two shots. It was a great experience for us to come up and experiment with different models and finally achieve an accuracy of 85%.
At Algoscale, we have a young dedicated team working with state-of-the-art deep learning algorithms and thriving to produce the best solutions with the latest technologies.
