Xview
Introductions
Among the multiple trends which COVID-19 has accelerated, digital fitness has received the most attention by media, companies, investors, coaches and athletes at all levels. Those fitness studios, boutique or otherwise, which were most adept at pivoting to and quickly expanding on their digital infrastructure survived while those who didn’t either closed their doors or were absorbed into larger players entirely. At the same time, the explosion of wearables — from smart watches to smart clothing — has transformed how movement is translated into data and data into actionable intelligence to improve, strengthen and win.
However, many of these systems provide “good enough” analytics for general improvement or bear costs which are prohibitively expensive for the majority of people. Professional coaches — albeit remotely available — are still required for quality and consistent feedback for most people. This is despite the fact that the complementary evolution of cameras on smart phones and motion capture technology has continued unabated but practically separate from the digital fitness platforms proliferating the android and iOS app stores today.
For example, while the core game of golf has not changed much since its origins in 15th century Scotland, understanding the game–how it’s played, how to improve, what makes the perfect swing for any player–has evolved significantly. Just as Moneyball ushered in the field of numbers to baseball and birthed the field of SABRmetrics, so too has AI helped make motion meaningful through radar, sensors and powerful camera systems and the latest mark of evolution: real-time, markerless AI in the game of golf.
Enter Xview
Leveraging a smart phone or ipad on the golf course has been done for years, but typically this analysis involves a post-game sit-down with a coach reviewing the video after the fact, manually drawing lines around specific points of the body and using the video to assess specific moments of play. Such analysis can happen hours after the game itself. In addition, the exercise is largely manual, time-intensive and while helpful for the student does not adequately scale to capture every swing.
To support the nearly 16,000 golf coaches throughout the United States, Xview has developed a system to provide real-time analytics so that both coach and student can quantify their movement together with automated tools to either highlight specific biomechanics or track a specific set of movements across a nine-stage analysis of the swing. An N=1 philosophy drives Xview–the idea that the optimal swing is specific to an individual’s strengths and weaknesses, body type or condition. The app, entirely free to download, brings structure to a coach’s lesson plan while also supporting recommendations, comments or concerns with empirical data. Furthermore, the digital platform helps coaches grow their business to both engages students at scale while finding new students to bring on, both near and far. Our aim is to build the most accurate and user-friendly golf instruction app on the planet.
Exploring Xview
Taking a full swing analysis provides a clean (and, not to be too humble, beautiful) visualization of a baseline golf swing and the user’s existing golf swing. Additional analysis can be accessed on the right-hand side.
Xview leverages trained models which can capture both the golf swing and biomechanics of the players. This allows us to analyze and automatically detect a full golf swing (with the potential to identify poor habits or potential drivers of injury).
Xview will start simply by enabling coaches and students to either live record their swing or upload from their phone a recorded swing to analyze.
Even absent a golf club, Xview’s position recognition enables the app to identify the correct position of the swing. Augmented positioning further allows a comparison between swings. Students and coachers, alike, will have the ability to additionally move forward and backward in time to identify minute movements of the swing.
Xview will integrate Ideas Lab’s revolutionary golf head analysis technology (integrating both golf club and biomechanics of the player — in the below case, Tiger Woods). Golf swing recognition allows us to categorize each position of the swing and proper placement of the hands and body.
Xview will include multiple folders to enable users to organize all videos (including both students and their own swings).
Annotation is a key tool for golf coaches worldwide and Xview was designed with the engagement needs of both coaches and students in mind. Multiple annotation tools will enable coaches on the app to directly analyze a given golf swing (with further means of communicating analysis of the golf swing to an individual or group of students — or other professionals agreed by the student, such as a physical therapist or fellow coach).
As shown in the following screen, our application would allow specific annotation for users, focus, for example, on hip placement and comparative body positioning to previous swings. Over the next several months, we will be looking to launch multiple annotation capabilities for golf coaches to analyze a student’s swing.
Close-up analysis would enable coaches to additionally track specific body points, such as the head. You’ll noticed also the 75 degree highlighted automatically.
At Xview, we see power as a math problem. Our aim is to help coaches and students understand how precise changes in movement can translate into a better game, less chance for injury, and ultimately a more powerful trajectory in one’s evolution from amateur to professional — no matter where in their athletic journey they begin working with Xview. Our focus on not only the numbers but also the experience and usability will help differentiate us from the handful of motion capture-based golf analytics programs out there. Indeed, other players in the market applying “markerless AI” to the game of golf have only validated the need of and demand for such products in upping the game of golf professionals everywhere. At the same time, beginners benefit by establishing the right form early on (before bad habits set in).
The Xview system will allow users to achieve professional-level analysis of their movement regardless of sport, level or objective. By working with biomechanics and sports coaches, Xview will capture the core analytics considered by experts when measuring movement, whether the objective is to strengthen the body, reduce chance of injury, or perform specific physical feats (run faster, ju
Analysis can be shared in one’s social network or with a select team of experts. Analysis can include the athlete against him or herself (e.g., comparing a motion done three months apart) or against professional benchmarks (either specific athletes or an average of multiple athletes in a given sport). Ultimately, Xview was launched to put the power of advanced analytics in the hands of anyone, anywhere.
How XView Differs
Our system differs not only in usability but also personalization, range of capabilities (i.e., multiple applications of markerless AI) and above all continual validation on both experience, accuracy and application. From Day 1 we have developed SME boards composed of golf professionals to help us understand what movement matters (for whom, in what context) and how best to visualize and illuminate this information to benefit both athlete and coach, alike.
Core principles of the Xview product will include:
- Goal-focused. Xview will be centered around specific objectives of the golfer with clear direction by the golf coach, community or self-defined goals. Objective and measurable output will be part and parcel of the Xview experience.
- Immediate. Xview will offer immediate, real-time feedback with each swing. Why real-time? In part because golfers will think about specific changes that should be made throughout the swing (swing thoughts) that would be lost if viewed as a video at a later time. Rela-time also increases both experience and meaningful engagement with the Xview system.
- Granular. Xview will provide a granular breakdown of the golf swing across 8 or 9 clear phases, each with specific body and club positioning. This granular assessment will provide both a more accurate and more meaningful feedback as they analyze their swing.
- Personalized. Xview believes that each player is distinct and any instruction should not be a one-size-fits-all proposition. This N=1 philosophy drivers our approach to player analytics.
Specific elements of our platform would include:
- Registration. Upon registering for the site, users will note their level, list their goals and objectives for joining the platform (e.g., join professional team, apply for scholarship, improve specific weakness, learn a new sport efficiently, work with a coach, etc.). This registration would create the benchmark or baseline for that coach or student’s continued use of the platform, with the more engagement yielding more actionable insights and goal-oriented recommendations. We are also working on an automated analysis of the player using our AI, or, new users can work with coaches on the platform for a more in-depth assessment of strengths & weaknesses (“Automated Coach”). The registration on the coach side will be centered more on supporting digital businesses or establishing the coach’s bone fides for the development of their coaching practices.
- Video libraries. Both students and coaches can upload videos onto the platform, making the videos either private (i.e., only those who ‘follow’ pages could see the videos) or public, searchable based on tags, say. Our Ai will be able to recommend training videos based on weaknesses identified through our system (e.g., tips to keep you hips correct throughout a golf swing). Users can see public videos of others at, or slightly above their levels, to improve their movement, and — as part of our product differentiation — comparisons can automatically be generated between a user’s motion and that shown in any selected video. Additionally, quantifying the motion between two (or more) videos can then be targeted to areas where there are greater deltas between a professional and novice, identified in our model through a combination of mathematics, AI, and domain expertise. Top videos, based on viewership, ratings, and relevance to the user, will percolate to the top, based entirely on a personalized video program.
- Upload & Analysis. Users can upload videos to enable our models to analyze and render each video, or frame, to become a medium for analysis. With each video, users can analyze their specific biomechanics, annotate specific areas of concern (our system could additionally build an automated highlight reel of “Moments of Concern” where the user’s biomechanics varies at some specific threshold from a professional average, say), share frames or clipped moments with the community, or view metrics specific to that motion. The actual analysis, together with any visualization for single analysis or in a dashboard layout, can be shared for collaborative analysis. Additional levels of analysis can be added based on the specific model or type of technology we’re applying in analyzing each movement or sport context.
Where we started
Xview is a product of Ideas Lab, a software engineering firm whose roots began in golf, led by Winston Yang, CEO, founder and lifelong golfer. For the past several years we’ve put technology development before productization, building up a team of AI, mobile development, UX/UI and other specialist engineers to help both the vision and execution of bringing the Xview product, when ready, to market. That day is today!
Additionally, we’ve also engaged with domain-specific Subject Matter Experts, who support us in translating the numbers in our app into meaningful instruction for the student at any level. If interested in joining our advisory board, contact Jesse at jesse@ideaslab.com.
Since our founding in 2017, Ideas Lab has been focused on building the world’s best markerless AI to help democratize access to powerful, accurate and professional analytics. Our initial focus on golf — a stick-based sport with limited but technical biomechanics throughout a golf swing — has expanded to incorporate multi-player sports with a wider spectrum of motion, such as hockey, basketball and yoga. Each individual technology we’ve developed — stromotion, biomechanics analysis, a combined stick-and-body pose model — has, to date, been developed in isolation.
Join us as we help bring forward the next evolution in golf analytics. Sign up here or contact us today at hellp@xview.ai.
ABOUT IDEAS LAB
Ideas Lab is an innovation lab and start-up studio building proprietary artificial intelligence, machine vision, and human motion analysis technologies. Today, while developing a suite of AI-based solutions, we are building a network of corporate and academic partners with whom together we will improve human performance in the many ways people move, perform and play!
Visit us to learn more about Ideas Lab today!
