CapstoneEngineerMagazineFall2026 - Flipbook - Page 26
BELOW: Dr. Noorbakhsh Amiri Golilarz and UA baseball
director of player development and recruiting Garrett
Wood stand on the mound.
E
VERY PITCHER WHO TAKES THE MOUND DURING
a baseball game carries with them their own distinct
pitching style. Within every delivery are subtle movement
patterns that re昀氀ect how the body coordinates throughout
the pitching motion. Understanding these patterns has
traditionally relied on the trained eye of coaches and countless
hours of video review. Today, advances in arti昀椀cial intelligence
are making it possible to analyze those movements automatically,
providing new insights that can support player development and
performance.
Dr. Noorbakhsh Amiri Golilarz, an assistant professor in
the department of computer science, collaborated with the UA
baseball coaching staff to create an AI-based app that provides
the staff with feedback they can use to better address their
pitchers’ weaknesses.
Golilarz serves as director of the Bioinspired Robotics, AI,
Imaging and Neurocognitive Systems (BRAINS) Laboratory, which
focuses on integrating principles from AI, cognitive neuroscience
and other technologies to create intelligent and adaptive systems
that move beyond AI toward human-like thinking.
The project originated when Golilarz received a call from
Kasen Rainey, director of pitching development and analytics
for the Alabama baseball team. Rainey had seen similar software
used in the MLB and was interested in developing one for the team.
“I knew this was something that was happening in professional
baseball, so I came to Dr. Golilarz and asked if he would be able
to build this type of system that is able to track segments of a
pitcher’s body,” Rainey said. “Our goal was to better understand
pitching mechanics and provide objective information that can
support player development and coaching.”
Golilarz’s research sits at the intersection of sports analytics,
biomechanics and arti昀椀cial intelligence, using computer
vision and machine learning models to analyze pitchers’ body
movements from broadcast baseball video, tracking the motion
of various body parts throughout the pitch.
“The system 昀椀rst detects the pitcher’s body joints — such as
the shoulders, elbows, hips, knees and arms — frame by frame
using pose estimation technology,” Golilarz said. “We then
analyze how these movements evolve over time before the ball
is released.”
The researchers 昀椀rst have to remove any noise, like other
players in the frame with the pitcher or items occluding the
pitcher from view, before the model can complete its analysis.
The resulting pose sequence allows the AI to identify subtle
movement patterns across the pitching motion, giving coaches
objective measurements that complement traditional video
review.
Rather than relying solely on manual observation, the system
uses machine learning to identify important phases of a pitcher’s
motion and highlight patterns that may warrant further review
by the coaching staff.
26 | CAPSTONE ENGINEER | FALL 2026
Dr. Noorbakhsh Amiri
Golilarz explains how
the AI system tracks
the motion of a player’s
body to analyze their
movement patterns.
