AI-powered app for measuring/tracking jump metrics
- Thread starter gsk8
- Start date
- Joined
- Oct 22, 2022
For training, maybe, as long as the coach wasn't trying to make all their pupils conform to one style. Some skaters naturally jump high, some jump far. I heard a commentator recently say a pairs girl was consistently landing throws on too bent a knee. Well, sorry, but I also always land jumps and throws on a deeply bent knee. I did it in gymnastics as well. It's my natural style, and I think the cushioning effect has preserved my knee joints for a long time.
If coaches started using this as a cookie cutter, skaters whose natural style doesn't fit the coach's measurements pattern will start to drop out with injuries. If the coach can re-calibrate for each pupil to use as a record of improvement within the skater's own style, then yes. To use it as a "one size fits all" pattern, no.
- Joined
- Jun 21, 2003
This sounds promising to me. Providing feedback for training purposes is a more realistic goal for AI than in judging. I don't see any harm in helping students and coaches keep a log of parameters like jump height, distance and speed of revolution. It's cheep, too.
Here's an article about the designer, Jerry Lu, an MIT graduate student.
news.mit.edu
Here's an article about the designer, Jerry Lu, an MIT graduate student.
Building a playbook for elite-level sports
A former competitive swimmer, MIT graduate student Jerry Lu works to optimize the performance of elite athletes and their sporting organizations. By developing his analytical skills through the master of finance program at MIT Sloan, he hopes to help build up smaller sports.
news.mit.edu
Widespread use in training could help refine it. If it is reliable (or as its reliability increases), it could eventually be used in judging.
I'm waiting for the user feedback. We should get some from figure skating bloggers and vloggers soon 
I guess the reliability of this app can be easily checked if you film the same jump with two phones and compare the data.
I don't expect wonders early on. Most likely, this technology will need tuning and further development. At the beginning, it will be rather a digital toy but it will be interesting regardless.
In perspective though, I think that it may evolve into an irreplaceable training tool. It can't teach how to jump but monitoring the metrics may help managing the overall physical form which is intrinsic for professional athletes who need to reach their peak form in time for important competitions. It could also inspire beginners who love to watch themselves improving
I guess the reliability of this app can be easily checked if you film the same jump with two phones and compare the data.
I don't expect wonders early on. Most likely, this technology will need tuning and further development. At the beginning, it will be rather a digital toy but it will be interesting regardless.
In perspective though, I think that it may evolve into an irreplaceable training tool. It can't teach how to jump but monitoring the metrics may help managing the overall physical form which is intrinsic for professional athletes who need to reach their peak form in time for important competitions. It could also inspire beginners who love to watch themselves improving
- Joined
- Jun 21, 2003
OK, but is this anything more than, for instance, a monitor than keeps track of how many steps you take every day when you go out jogging? If you know that yesterday you jumped 60 cm into the air and today you jumped 62 cm -- where dies the artificial intelligence part come in?
If you know that you normally jump 65 cm and then you do all the same and you see it's 54-56 cm and you know what you ate yesterday evening...
- Joined
- Dec 9, 2019
How can it be? I understand that deeply bent knees are considered best, and it makes sense as it protects hip and knee articulations, and requires precision in controlling the energy transfer? Now I'm worrying: are there coaches who ask their students NOT to bend their knees at landing? Could the reason why the "big jumpers" of the last decade had/have this problem of stiff landing, which seems quite bad for their hips?
- Joined
- Dec 9, 2019
This is also what I'm wondering, the level of technology. As far as jump measurement goes, I don't see a need for AI, particularly during a testing phase? Having a data memory for each skater and many graphs possible with these would be fine but still not AI?
As I have a thwarted mind, after the first "yay", I thought, not all coaches will adopt it, particularly those who have their students believe wrongly that their jumps are good...
- Joined
- Aug 8, 2023
I will be impressed if this proves to actually measure "jump height" consistently with only an iPhone camera. This all sounds ridiculous to me.
Valid question in any context regarding "AI" news.
- Joined
- Oct 22, 2022
I don't remember which skater at what competition, but I just assumed it was her natural style and perhaps she'd fallen a couple of times and the commentator was coming up with the first reason that came into his mind. As far as I know, bent knees are your shock absorbers and are good. Occasionally in throws I don't get my usual height, and the ice and I meet up sooner than expected. There's quite a jar travels up through the leg and spine if my knee hasn't had time to bend enough.
I doubt if it's a coaching thing, to ask the skater to deliberately land stiff-kneed. I'd assume what they needed to work on was the height of the jump, so perhaps this new coaching tool would help with correcting that.
Doesn't it say clearly that it uses Computer Vision to come up with all the stats and landing quality?
Computer Vision (AI overview):
"Computer Vision is a field of AI that trains computers to interpret and understand visual information from images and videos, enabling them to "see" and act on what they perceive. It uses machine learning, deep learning, and other techniques to perform tasks like object recognition, facial recognition, and pattern detection, powering applications in self-driving cars, medical imaging, and manufacturing defect detection.
How it works:
I understand that the general principle is similar to that of Ice Scope. Except that Ice Scope needed a human operator but in this app the function of human operator is given to AI.
"Computer Vision is a field of AI that trains computers to interpret and understand visual information from images and videos, enabling them to "see" and act on what they perceive. It uses machine learning, deep learning, and other techniques to perform tasks like object recognition, facial recognition, and pattern detection, powering applications in self-driving cars, medical imaging, and manufacturing defect detection.
How it works:
- Image conversion: A computer converts a digital image into a grid of pixels, with each pixel represented by numerical values for its color and brightness.
- Pattern recognition: Algorithms use machine learning to find patterns in this numerical data, which helps them identify and classify objects or features within the image.
- Analysis and action: Based on the analysis, the system can identify objects, detect anomalies, or take other actions, just as the human eye and brain do."
I understand that the general principle is similar to that of Ice Scope. Except that Ice Scope needed a human operator but in this app the function of human operator is given to AI.
Last edited:
- Joined
- Jun 21, 2003
About computer vision, here's an interesting historical factoid. A pioneer in this field, recognized in retrospect, was Florence Nightingale, the famous nurse in the Crimean War (1850s) who was also the first woman elected to the Royal Statistical Society. She invented the circular histogram for representing digital data visually, a technique that robotics people relied on in the 1990s to teach a robot how to move about without running into objects in its path.
Last edited:
- Joined
- Jan 31, 2019
Lol, the skaters' inboxes and messaging apps will soon be overflowing with, "Your skating keeps me awake at night, it's so beautiful. You deserve so much more in GoE! Please, contact me via telegram to discuss some ideas on how to make judges recognize your greatness!!!"
I dunno if anything gets this, but as a writer I am continuously spammed by stuff like that from the 'digital artists'. Scammers love AI and the suckers who don't know how to use it to achieve their own ends. As a minimum, each skater needs to know what it is, and what it can do to defend against the bots.
I dunno if anything gets this, but as a writer I am continuously spammed by stuff like that from the 'digital artists'. Scammers love AI and the suckers who don't know how to use it to achieve their own ends. As a minimum, each skater needs to know what it is, and what it can do to defend against the bots.
- Joined
- Aug 8, 2023
Hahaha, this is so true! I wish I could give you multiple laugh reacts.
Very astute connection, I didn't even think about this.You're definitely right! It's especially funny to anyone who was a precocious business-person in their teens and mercilessly solicited basic services.
I imagine it's far worse today. The effort of searching directories, expos or generally networking is completely lost. People must receive hundreds of messages. Well—my brief experience with LinkedIn also suggests this. Living without social medias or open online-inboxes is paradise.
Last edited:
- Joined
- Jun 21, 2003
My favorite is, "I feel so human when I talk to you."
- Joined
- Oct 22, 2022
Even I've got a couple of those, and my partner and I just perform at our club. We toy with the idea of adult competition, but not with enough enthusiasm to actually do it. I guess our names are picked out of the club's newsletter and various posts, but when I see remarks about how "the judging is so unfair" to us and how this ersatz person could help improve our style, I just laugh and delete.
(I get them as a writer/storyteller too. Do they offer to be your agent and help you find a publisher, or to self-publish? Thanks, but no thanks even if they were legit. Been there, done that in the newspaper world, for too many years in the past.)
- Joined
- Jun 21, 2003
By the way, I do have to say that as far as figure skating judging is concerned, I think that the whole AI discussion is exactly backwards. I don't think that there is any useful role for AI in determining techy things like the height of a jump, speed across the ice, or degrees of q-ness.
Where AI could possibly have a contribution to make is on the esthetic, PCS side.
The original goal of AI was to built a robot that can "think" like a human -- or, since we have little understanding of what "think like a human" means, at least to mimic human use of language and action so well that we can't tell whether we are talking to a human or to a robot.
Facial recognition is mentioned above (post 13). We humans have some sort of intuitive build-in talent for doing this. We can usually recognize the faces of our friends and often of more distantly-connected people that we have seen pictures of. How do we do it? Damfino. But yes, a computer algorithm that can analyze pictures of human faces and say, "Yup, that's Donalf Trump, all right, bless his heart" -- that might fool us into thinking that the computer is human, or at least that it is performing a task that is in the bailiwick of human intellectual capacity.
One can imagine extending this capacity into judging music contest or poetry contests... or figure skating contests. The self-learning computer could analyze musical structure, or examples of of word usage that makes us humans say -- "yes, indeed, that's a poem all right, and a pretty good one". For figure skating an AI algorithm could be trained to "watch" a figure skating performance and say, that is the kind of performance that hunman figure skating judges tend to give 9,25 in Choreography for."
Whether this would be a good thing or bad, 't can't say. To answer that question I would need an AI algorithm trained to make moral judgments like humans do.
Perhaps the computer would say, before I tell you what score this program gets in Composition, first tell me nationalities of the judges."
Where AI could possibly have a contribution to make is on the esthetic, PCS side.
The original goal of AI was to built a robot that can "think" like a human -- or, since we have little understanding of what "think like a human" means, at least to mimic human use of language and action so well that we can't tell whether we are talking to a human or to a robot.
Facial recognition is mentioned above (post 13). We humans have some sort of intuitive build-in talent for doing this. We can usually recognize the faces of our friends and often of more distantly-connected people that we have seen pictures of. How do we do it? Damfino. But yes, a computer algorithm that can analyze pictures of human faces and say, "Yup, that's Donalf Trump, all right, bless his heart" -- that might fool us into thinking that the computer is human, or at least that it is performing a task that is in the bailiwick of human intellectual capacity.
One can imagine extending this capacity into judging music contest or poetry contests... or figure skating contests. The self-learning computer could analyze musical structure, or examples of of word usage that makes us humans say -- "yes, indeed, that's a poem all right, and a pretty good one". For figure skating an AI algorithm could be trained to "watch" a figure skating performance and say, that is the kind of performance that hunman figure skating judges tend to give 9,25 in Choreography for."
Whether this would be a good thing or bad, 't can't say. To answer that question I would need an AI algorithm trained to make moral judgments like humans do.
Perhaps the computer would say, before I tell you what score this program gets in Composition, first tell me nationalities of the judges."
Last edited:
- Joined
- Jun 21, 2003
That does illustrate one sad fact, though. From a marketing point of view the one universal constant that can be relied on is that all figure skaters automatically think that the judges are unfair to them, and thus are good candidates to be fished for.
Similar threads
- Sticky
