top of page

How SmartPitch came to be

Dexter is our son, a lefty pitcher who has been passionate about baseball since he was 8 years old. He pitched in Little League, JV, High School, college and now men's leagues. Dexter is also a wonderfully talented geek, a lead  firmware & embedded engineer who loves designing & building connected devices and Internet of Things (IoT) devices. Back in 2009 we were counting frames in videos of his pitching to measure speed. By 2013 we had built programs to speed up this manual counting and speed measurement.

​

I am a retired data scientist with a PhD in Quantitative Business Analysis, and also love doing baseball things with Dexter. That includes being his bullpen catcher, and his partner in co-creating SmartPitch. Our baseball connection goes way back to when he fell in love with the sport when he was 8 years old.

​

He wrote a high school essay about his first time at a major league baseball game when he was 8, at Fenway Park.  It ends with: "The important part was that my dad and I had discovered a new passion, something that would bring us endless joy and fun in the future!"

 

More than 10 years later, together we discovered that studying his delivery in detail, frame-by-frame in those videos was a big help in improving his delivery. A by-product was that we also developed a way to measure his pitch speed.

 

In the early days of developing our speed calculations methods, we measured the ball's travel in each video frame on a desktop computer, using manually operated precision scale calibration and measurement software.

​

But by 2014, as Dexter was graduating from Stony Brook University with a degree in computer science, he came up with a critical breakthrough.

​

He created ingenious computer code that automated the video image capture and analysis of video files of his pitching sessions. Our testing of the outputs from his code showed great promise in its ability to identify the ball in the video image live and in real time, while filtering out the noise of movement in the background. It also proved the accuracy of the data that his code captured from the video.

​

The raw data from the video analysis becomes the input to the math algorithms that I wrote and fine-tuned. These mathematical factors allow you to use SmartPitch on a standard smart phone from many locations on the field, including from the dugout, in foul territory or in the stands. This is a breakthrough versus radar guns that force you to stand behind home plate.

SmartPitch is far more useful in real-life baseball situations than other apps that must be placed on the infield and can't be used during a ball game.

​

It is also better than radar.  You cannot measure hitting launch angle or exit velocity with radar. And you must be positioned exactly behind the catcher or pitcher, no where else.

​

The final component required was the smartphone user interface (UI). We designed it based on many experiments and trial and error and came up with the minimum number of user inputs for really fast set-up, which takes about 25 seconds.

Knee Movement Analysis.PNG

Analysis of Effect of Knee Angle on Speed 2013

That discovery confirmed the viability of one of the three essential components for the SmartPitch app - the math I had developed. It also fired us up with the possibility that we could build a smartphone app based on our method for determining speed very precisely from live, automated video analysis of pitching.

​

The "big moment" came when Dexter did the tricky C++ / Objective C coding to combine these three components: the UI + the live video feed capture and analysis + the math algorithms.

​

Now the crucial test remained: would it run fast enough to work on a typical smartphone in real time?

Frame-by-frame.PNG

Frame-by-frame Ball Trajectory from Patent Application

In bullpen sessions, a pitcher throws a pitch every 15 - 20 seconds. Also, the app uses the live video feed from the smartphone camera. A new video frame is created every 1/30th of a second, and the video analysis software had to be faster than that analyzing each frame, not to fall behind the pace of the constantly incoming streaming video data.

​

The major challenge was to assure that the thousands of lines of code and math analysis we had written had to finish processing in less than 1/30th of a second.

​

So that April when we loaded that combined code onto an iPhone 5s, after much debugging, we had the enormous father and

son thrill of celebrating years of working on our method together.

It worked!!!!

​

We had our essential breakthrough. We tested it rigorously with live pitching and proved that it analyzes the real time streaming video feed with precision, and fast enough to work in live pitching situations.

​

That was the green light we needed for our smartphone speed gun Proof of Concept (POC) hands-free app you can use almost anywhere on the field, SmartPitch!!!!

​

 -- Chuck Richard, PhD, SmartPitch CEO & Cofounder, and Dexter's Dad

bottom of page