The calls to Kolkata kept coming. Steven Cadavid, the CEO of biomechanics analysis company KinaTrax, would reach out to his data-management partner, iMerit, last October to say that the ongoing project might last only one more week. Then there’d be another week extension. And another.
“It’s funny, we didn’t even know, initially, that we were doing this work for the Cubs,” iMerit vice president of technology and marketing Jai Natarajan told SportTechie, noting that the progress of last year’s Major League Baseball postseason wasn’t a newsy topic in eastern India. “It wasn’t until much later that we realized, the Cubs just kept winning. We didn’t connect the dots.”
This season’s Chicago Cubs’ playoff run ended last week and hardly went unnoticed. The World Series title in 2016 sparked a legion of fans within iMerit. Cadavid shipped Cubs hats and sent articles about baseball and about KinaTrax (including one at SportTechie) via Slack to the novice fans.
What KinaTrax offers baseball clubs is markerless motion capture and computer vision to replicate in game settings — with appropriate adrenaline, stress and tension — what is usually relegated to the controlled environments of a biomechanics lab. These assessments are used for injury prevention and performance enhancement, whether it’s identifying variance in internal and external shoulder rotation or improving mechanics to generate more velocity or avoiding subtle tendencies of pitch tipping.
KinaTrax’s next definite plan is to delve into hitting analysis — particularly the tracking of the bat’s motion through the strike zone — and one other proposed area of expansion is near and dear to many of iMerit’s workers: cricket.
Cadavid and Natarajan both said conversations in the cricket world have been exploratory but encouraging thus far. Natarajan noted some initial interest given a spike in injuries to cricket bowlers in the last decade, primarily back and shoulder from pace bowlers. (Bowlers are the sport’s counterpart to baseball pitchers, although the motion prohibits any bend to the elbow; pace bowlers essentially throw fastballs, while spin bowlers throw exclusively breaking balls.) The motion-capture technology could have the same application as in baseball — reducing injuries, honing mechanics — but also for detecting possible elbow flexion violations.
“Pace bowling is just thankless,” Natarajan said. “You’re just putting your body through nasty contortions.”
In fact, Natarajan is so bullish on the technology and its potential in discerning discrete examples of “explosive human actions” that Cadavid laughs when he hears how many sports to which his colleague wants to apply the technology. “I need Jai as a rep,” he quipped. Natarajan said a tennis serve would be another good example, as well as a golf swing. “In golf, the dollars are concentrated into that one movement, which is the swing,” he said, before adding with a laugh: “Everything else is walking around, right?”
KinaTrax may not be relegated to capturing isolated movements any longer, either. Cadavid said the technology exists for multi-person tracking through wide-angle, high-resolution cameras that would open up opportunities for tracking in the NBA or NFL.
For now, KinaTrax extracts its raw footage from a standard set-up of 12 high-frame cameras trained on the pitcher’s mound. The video is sent to iMerit, where employees identify snippets that will yield good visual data and then annotate joint positions and limb angles. These markups are returned to KinaTrax, where its machine learning algorithms stitch together 3D biomechanical models. (KinaTrax has three MLB clients: the Cubs, Tampa Bay Rays and one other whose agreement includes confidentiality.)
The 50 iMerit employees who do work for KinaTrax have gone “from completely oblivious to baseball to now where they could probably be pretty decent at picking out good pitchers,” Cadavid said. “They look at literally thousands of images of slow-motion pitches.”
With such rapidly accrued granular insights, several are now making recommendations to Cadavid, telling him, “Hey, this pitcher looks pretty good to me.”
Such insights would have seemed impossible to iMerit’s employees five years ago, not only because of the geographic and cultural distance but also because of the socioeconomic gap. iMerit is a social-impact for-profit company employing many in India who otherwise would not have access to white-collar professions in a field such as data analysis. Of its more than 1,000 employees, iMerit draws about 80 percent from really poor backgrounds and 55 percent who are women — both anomalies in the tech sector. They are educated but still had difficulty advancing.
“Our goal is to create digital livelihoods for underprivileged youngsters,” Natarajan said.
Former Hewlett Packard general manager Radha Basu founded the company in 2012 and received its early funding by three leading impact funds: Omidyar Network (started by eBay founder, Pierre Omidyar), Khosla Impact (begun by Sun Microsystems founder Vinod Khosla) and the Michael and Susan Dell Foundation. Some 80 percent of iMerit’s employees are graduates of a training program run by the Anudip Foundation, which Basu and her husband, former Cisco executive Dipak Basu, founded in 2005.
iMerit believes AI and automation should create jobs, not eliminate them, which Natarajan acknowledges sounds “counterintuitive.” He added that humans can augment the work of intelligent algorithms, and that iMerit seeks to staff that sector of the data ecosystem. (Among its other project is working to help train the AI that powers self-driving cars.)
“There are some really interesting dynamics where some of the most pioneering tech people are really trying to figure out how to bring people back into the equation,” said Natarajan, who is also a board member at the Anudip Foundation, adding of the prospect of eliminating the role of humans altogether: “It’s not feasible and it’s not desirable either.”
Submit Your Nominations For The SportTechie Awards!
Now, iMerit has been a part of KinaTrax’s team since the company launched in earnest for the 2016 season. Cadavid’s own journey to baseball was roundabout: a computer engineering Ph.D. at Miami, he worked with developmental psychology professor Daniel Messinger on a pioneering project to make early autism diagnoses by using computer vision and machine leaning to quantify movement and social interactions a between mother and baby in a controlled environment. Cadavid, who co-author several journal papers, gained an interest in markerless mo-cap field. His first project, nuiCapture, boosted him high enough on Google’s search results that a group looking to pursue a baseball project contacted him. His interest in the sport had previously been minimal — “Not much,” Cadavid said with a laugh, “I am now” — but accelerated along with his international business partners, for whom this undertaking has been nothing short of extraordinary.
“Think of the incredible transformational power of somebody from a really underprivileged background sitting in a small-services company like ours and actually being a part — a very small part, but a tangible part —of the Chicago Cubs winning the Series after 108 years,” Natarajan said. “Think of the journey there, what that does for them psychologically. They feel a part of this modern world. They don’t feel left behind anymore.”