Pixel Machha
Photos from Camera to your Guests in lightening speed !
234 followers
Photos from Camera to your Guests in lightening speed !
234 followers
Pixel Machha gives event photographers and social media a professional platform — FTP camera uploads, AI face recognition, branded guest albums, Social Media specific editing and exporting, and live slideshows. All in one place.
@sumanth_bettadapura It's a wonderful product. This is very useful for me as i can get all the photos i am in easily while the event happens especially with the selfie feature .
How does the face recognition handle group shots with people who haven't opted in, and is the AI matching done on your servers or locally on the photographer's machine?
@fatmaodef39623 I handles the group shots very well, I mean for people who hasn't opted in wont be able to find their photo but the others who has opted in they can find their photo even if they are in group photo,. The AI matching happens over server side, dont want to put pressure over the photographer's machine and also its something not possible to have computation over cameras. cameras dont have that ability.
the FTP upload straight to guest albums is a genuinely nice detail, most tools in this space make you export from lightroom or your camera app first which kills the "live" part of live events. at a big wedding with a few hundred guests, does the face recognition hold up with people in and out of shadow/backlight, or does accuracy drop enough that photographers still have to manually sort some of it
99% is solid if that holds in low light too, wedding dance floors and dim receptions are usually where face matching falls apart. is that number from indoor venue testing or mostly daytime outdoor stuff
@omri_ben_shoham1 Want to be honest here, I said 99% with the experience I had from using the platform, for face recognition I am using AWS service called Rekognition and Its super accurate, I have not built the face recognition from the ground up, I'm just using AWS Service for that, but that's a very solid service - I dont think I could have done better job myself compared to their service.
honestly that's the smart call, no point reinventing face recognition when Rekognition already handles it well. curious about the cost side though - are you paying AWS per API call as usage scales up, or did you lock in a flat rate? seems like that could get pricey fast with photo-heavy events like weddings.
$223 for 2 days is rough, especially if it's mostly architecture overhead and not actual usage. before you build your own face recognition from scratch, worth checking if it's a data transfer / instance sizing issue rather than the recognition calls themselves - that's usually where AWS bills balloon on short events
@omri_ben_shoham1 what actually happened is a missed bug, on each selfie upload I was scanning through all the photos via Rekognition Face search feature , What I had in mind was keeping the vector data with myself and comparing the vector data of the selfie against the pre parsed photos, since that didn't happen as rekognition was not providing the vector data the bill shot up, hence Im looking at implementing face recognition system of my own, that too in the sense I would be using Open Source available solution than building everything from scratch.
but thanks a lot for the concern and the interaction you are doing <3
daily.dev
This is an excellent solution to a big problem in any conferences… I have used it in recent React Nexus... It was amazing. All the best with it. 👍
@tapasadhikary Thanks a lot for your kind words and glad to hear that it came in handy for you
makes sense re: the vector data approach, that's a much cleaner design than re-scanning the whole gallery per selfie. open source face rec (insightface or similar) should get you most of the way there without the per-call AWS bill, just budget extra time for GPU inference if you're running it yourself instead of a managed API
Blaze
Congrats on the launch!
@akashhamirwasia Thank you Akash !
@akashhamirwasia Thanks a lot Akash !
Macaly
instant guest delivery is such a pain solved 📸 congrats on the launch!!