[HN Gopher] Show HN: Hooper - AI-driven stats and highlights for...
___________________________________________________________________
Show HN: Hooper - AI-driven stats and highlights for basketball
play
Hey everyone, OP here. Wanted to share a bit more about Hooper -- I
started building it with a good friend of mine six months ago. We
play a lot of pickup together and were arguing about who has a
better jump shot and ended up hacking together an app to settle it
The way Hooper works is you can record yourself using the app and
ideally a tripod (optional). The app will track everyone, whether
its a solo practice, a 3v3, or a 5v5. We think there's a lot of
stuff out there for basketball drills but what we really wanted
Hooper to be for is actual game play. That means, it can do things
like track multiple players, sync two half court recordings, and
differentiate 2s vs 3s. Once you finish recording, it'll process
for a bit and then ask you to tag yourself (and optionally other
players). Then it spits back out a few things: you can watch the
full footage as well as a clipped version that's only the
interesting plays; you can see highlights and offensive box stats
for every player. When you sign up, you get a Hooper profile that
tracks your overall stats across the sessions. You can also do
things like build a mixtape from your highlights for insta. You can
friend other players on Hooper to see their profile & comment on
their games (also add us "grub" and "kangexpress"!) We've been in
a closed beta for about 3 months now, fixing bugs and getting
things to work with a set of early adopters. We are now starting
our open beta ! If anyone here wants to try it out, you can just
download the app on https://www.hooper.gg/?utm_campaign=h1. We are
early in our journey, and lots of improvements to be made in the
next few months, but we would love your feedback and ideas!
Author : grub007
Score : 100 points
Date : 2024-07-24 21:39 UTC (1 days ago)
(HTM) web link (www.hooper.gg)
(TXT) w3m dump (www.hooper.gg)
| elpakal wrote:
| This is a cool idea, nice work. I spent some time as an assistant
| college coach in the NCAA and the amount of time spent reviewing
| film and capturing key plays/schemes etc is huge. I do some
| computer vision and always wondered if it would ever be accurate
| enough to skip over dribbling up the courts, time outs etc so
| scouts could just get to the important parts of the film--it
| looks like you are on track with what you claim it can do. I
| might sign up for the beta... good luck!
| grub007 wrote:
| Yes please try it out, I'd love to know what you think! My
| email is mike@hooper.gg if you want to chat more about
| basketball and tech. I would be curious to learn what you have
| tried so far and what was important to you as a coach
| bl0b wrote:
| Awesome! I someday hope to find the time to make something like
| this for soccer - individual skill drills as well as gameplay
| analysis.
|
| Care to share any technical details about how your analysis
| pipeline works? :)
| grub007 wrote:
| Yeah, it would be super cool if all sports had things like
| this, i think it really enriches the player's experience. I
| know in tennis there is swingvision so definitely seems like
| things are headed that way.
|
| We ended up having to collect our own data and train our own
| deep learning models for detecting balls, players, court,
| actions, and tracking. But this was because we needed to make
| things work almost all the time to achieve our desired
| accuracy. You should try out off-the-shelf things like YOLO,
| Segment Anything, and Detectron and then after that you
| probably need to invest more on data collection and model
| training.
| imp0cat wrote:
| Adidas made some soccer balls with built-in sensors that can
| provide data in almost real-time to your phone (ie. take a
| shot and review ball speed, rotation, point of strike
| etc...). Video:
| https://www.youtube.com/watch?v=mpIOnnU1R_o&t=119s
| (unfortunately no english subtitles).
|
| Also, the technology was used in their latest Euro 2024 ball
| - the Fussballliebe - to help video referees and their
| decisions.
| marban wrote:
| What are you using for video analysis?
| grub007 wrote:
| Baked our own image models that process each frame to detect
| objects. Then we added some lightweight video models to detect
| actions and tracking
| drited wrote:
| Cool but offence is only half the story! Any plans to add
| defensive stats or stuff like turnovers / steals etc?
|
| Also any plans to make it work from camera footage e.g. A 360
| camera that can capture the whole court at once instead of
| syncing?
| grub007 wrote:
| Yeah! We are planning to add blocks, steals, and rebounds in
| the coming months. I think fouls is still a reach goal right
| now (hard even for humans) but maybe one day.
|
| Re: capturing the whole court, I think we are considering
| supporting panning of the camera (in addition to syncing). We
| focus pretty heavily on making sure things work with just a
| normal phone so it's more accessible to everyone. We are
| definitely starting bottom up but maybe one day we will get to
| more high tech hardware.
| jussy wrote:
| Re:fouls.
|
| At my pickups courts we call our own fouls by saying Foul.
| Perhaps the mic can pickup key words matched the play
| stopping.
| grub007 wrote:
| Oh neat, yeah we could some speech to text thing or maybe
| even some like visual signal thing to call fouls.
| drited wrote:
| Very cool of you to take on board feedback.
|
| In relation to user experience on first try...I just
| wanted to test it out for 30 second to see if it's worth
| keeping. I haven't tested now, will have to wait until I
| get an hour free later because there are a few
| roadblocks. - You require email and password registration
| and to click a verification link. Not too bad - When
| first launched the app asks you to allow mic permissions
| by clicking settings. However in Hooper settings on
| Android there is no mic permissions option. Bad. - You
| require 5gb free to use the product at all.
| Understandable given storage is required for video but
| how about reducing that so users can do like a 30-second
| test for first use? That way they don't have to spend an
| hour going through their phone's media to see what they
| want to keep and what they can delete.
|
| You'd be amazed at how many people (like me lol) have
| almost full storage on their device most of the time.
|
| Edit: oh....it's not a setting for mic permissions, it's
| a general permissions setting and if you touch that you
| can add additional permissions. Based on the text of the
| popup I hadn't guessed that was the right place to look
| and I hadn't realised it was possible to add permissions
| by touching that area of the settings menu.
| dvt wrote:
| Very awesome, congrats on your release! I know of a similar golf
| app that made absolute bank, so if there's a niche and you can
| tap into it, you guys will likely do quite well.
| grub007 wrote:
| Appreciate the kind words!
| potamic wrote:
| What's the golf app?
| grub007 wrote:
| Here are some i looked at - SwingU, Ace Trace, Swing Profile,
| Sportsbox AI. Overall seems like two kinds of golf apps: (1)
| more micro- to track and grade your swing, and (2) more
| macro- something to track your progress over a course
| jussy wrote:
| This is very interesting and well executed for the initial
| release. Good luck!
|
| One thing I found odd was that the default/main screen in the app
| is a feed. That feed is just short videos without any of the game
| data the app captures. Perhaps an overlay would showoff the app
| as well as the skills. The videos by themselves don't add much
| differentiated value.
| grub007 wrote:
| Yeah, that could be cool -- we played around with some designs
| on showing the leading scorer or some ranking in each feed card
| as well. I can't say we've cracked it yet but I like the idea
| of an overlay.
| foobarkey wrote:
| Well done, ai_bullish++
| Avisan wrote:
| This is awesome, thanks for sharing! Hooper sounds like a
| fantastic app for pickup basketball players. I love the idea of
| tracking actual gameplay and creating highlights. The ability to
| sync recordings and differentiate shots is impressive. Can't wait
| to try it out and see how it tracks my games. Great work!
| grub007 wrote:
| Yeah! Excited to see what you think of it. Making our tracking
| algos and shot classifiers better is something we think about a
| lot.
| frankdenbow wrote:
| Amazing. Pitched something like this to a friend a few days ago
| and its awesome that you've built it. Headed to a game now and
| down to use it and give feedback (im in two groups and both have
| people setting up phones to record). Now add in an AI ref so
| people can stop arguing for 10 minutes over wether someone
| stepped out of bounds or not in a pickup game.
| frankdenbow wrote:
| https://youtu.be/SUvfKX7yXLo?si=0C1XTGOa2NTUEaHy
| grub007 wrote:
| oh dang! Thanks for the feature- this was sick to watch.
| Appreciate you putting the time in to give your thoughts.
| frankdenbow wrote:
| Used it this morning and had a few quick thoughts:
|
| 1) Love the feed, realized its nice to see other random games i
| have no connection to. Wouldn't have thought it would be
| entertaining but its great.
|
| 2) Export video to camera roll: im sure some folks will want to
| export the whole thing (update: now i see that you can do it
| when you full screen it, may want to move that to the menu on
| the video processing screen).
|
| 3) Creating sharable clips that are branded can definitely help
| grow the app. Seems like you may be able to identify types of
| highlights, so mashing them up, with some graphic overlays and
| voiceovers (think play by play with elevenlabs). Have a few
| folks who work in this space for sports teams that may be able
| to give some guidance.
|
| 4) Turning the screen off while recording is a brilliant
| feature and is reassuring
|
| 5) Realtime scoring is likely harder but would be a killer
| feature.
|
| Love this, all the best on the development!
| grub007 wrote:
| This is awesome - thanks for trying it out and for the
| feedback! I really appreciate it and it goes a long way to
| help us make the app better.
|
| +1 on the graphic overlays and voiceovers. I personally want
| to add Mike Breen's Bang bang voice over as an option haha.
| Also check out mixtapes on your profile page if you haven't
| yet. You can mashup highlights that way from different
| recorded sessions.
|
| +1 on realtime scoring and processing. Right now, a recording
| takes ~30min to process. We are trying to reduce this over
| time as much as we can until it is close to live.
| thom wrote:
| 92% accuracy is really good! Is this just action recognition and
| player identity at the moment or are you actually generating full
| tracking data under the hood?
| apwell23 wrote:
| This looks great. curious, Is is something similar for tennis
| and/or ice skating?
| grub007 wrote:
| Check this product out - https://swing.tennis. I've used it
| before and it's quite good
| seanhunter wrote:
| This looks pretty sick and the site and promo video is very well
| done also. Nice job.
| grub007 wrote:
| Thanks! Appreciate the comment!
| goodmattg wrote:
| I love this! looked into doing a similar project, you're
| competing against Hudl but using the phone instead of custom
| hardware (always preferred). Highlight segmentation may be a
| challenge with SOTA cv methods, but there are lot of directions
| you can go in.
| grub007 wrote:
| Yeah Hudl is definitely a beast in the space though I think of
| them as almost more focused on orgs and teams rather than
| consumers.
|
| Re: highlight segmentation was and still is a challenge that we
| work on. In the beginning we had a hard time dealing with false
| positives when our models thought a shot was in but it wasn't.
| This has gotten better over time with more data and is hovering
| near 92% acc these days but obv not 100%. But i'm optimistic as
| data grows and the sota methods get better, this isn't the
| hardest ml problem in the world
| goochphd wrote:
| This combines three of my greatest passions - basketball,
| computer vision, and analytics. I love it! Thanks for sharing :)
| grub007 wrote:
| Thanks for the comment! Would love to know what you think and
| how we can make the app better :)
___________________________________________________________________
(page generated 2024-07-25 23:15 UTC)