Загрузка видео...

Не удалось загрузить видео

На главную

I'm testing the new Segment Anything Model on satellite imagery. It works so well! Within 30 seconds I can get the footprints of all the objects of interest in a scene. Colab notebook below Earth Genome

clkruse's profile picture

Caleb Kruse

1,168 subscribers

90,043 просмотров • 3 лет назад •via X (Twitter)

Наука и технологииИскусствоОбразование
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 10

Фото профиля Caleb Kruse
Caleb Kruse3 лет назад

Trying it on a number of different objects in Bali

Фото профиля Caleb Kruse
Caleb Kruse3 лет назад

Try the notebook here. Requires a Mapbox API key

Фото профиля Caleb Kruse
Caleb Kruse3 лет назад

you seem like a guy who could get this running in the browser

Фото профиля Kendall Hines
Kendall Hines3 лет назад

@EarthGenome Now measure the sq footage of that area and we got something!

Фото профиля Ricardo Yrupailla
Ricardo Yrupailla3 лет назад

@EarthGenome @rburhum

Фото профиля Anurag Sharma
Anurag Sharma3 лет назад

@EarthGenome That's awesome! May I know which dataset you used? Can the OSM data work?

Фото профиля Lawrence Griffiths
Lawrence Griffiths3 лет назад

@EarthGenome @JohnCassidyGB

Фото профиля Walid
Walid3 лет назад

@EarthGenome This is amazing! thanks for sharing!

Фото профиля James Beamish
James Beamish3 лет назад

@EarthGenome Can’t wait to try it, thanks for posting!

Фото профиля Iván Cárdenas
Iván Cárdenas3 лет назад

@EarthGenome @SaveToNotion #thread #GIS #remotesensing

Похожие видео

Interactive well pad identification in the Permian Basin using Meta's Segment Anything Model 3. Draw a box around one well pad using Mapbox imagery; SAM3 finds similar objects in the view. So many use-cases for AI in oil & gas, and I'm not just talking about ChatGPT.
0:41

Interactive well pad identification in the Permian Basin using Meta's Segment Anything Model 3. Draw a box around one well pad using Mapbox imagery; SAM3 finds similar objects in the view. So many use-cases for AI in oil & gas, and I'm not just talking about ChatGPT.

Kyle Walker

12,214 просмотров • 6 месяцев назад

Once the biggest iceberg on Earth, A23a has fractured in the Southern Ocean, shrinking to the size of Greater London in new satellite imagery.
0:23

Once the biggest iceberg on Earth, A23a has fractured in the Southern Ocean, shrinking to the size of Greater London in new satellite imagery.

AccuWeather

19,724 просмотров • 11 месяцев назад

✨ Introducing 3 new ways to build with real-world imagery and AI → From the fine details of a city street to a bird's eye view of the planet, at Cloud Next we revealed new ways that businesses can use our Street View, aerial and satellite imagery, and tap into our comprehensive view of the physical world including: 🏙️ Use Maps Imagery Grounding in Gemini Enterprise Agent Platform to generate location-specific assets at scale, anchored in real-world Street View imagery 🌍 Unlock a new perspective with Aerial and Satellite Insights, part of Earth AI, in BigQuery and Agent Platform to quickly extract detailed info about our changing planet 🛰️ Build custom apps with Aerial and Satellite models in Model Garden from Earth AI that easily identify specific objects or scenes Read the blog at the link above and sign up for early access.
0:20

✨ Introducing 3 new ways to build with real-world imagery and AI → From the fine details of a city street to a bird's eye view of the planet, at Cloud Next we revealed new ways that businesses can use our Street View, aerial and satellite imagery, and tap into our comprehensive view of the physical world including: 🏙️ Use Maps Imagery Grounding in Gemini Enterprise Agent Platform to generate location-specific assets at scale, anchored in real-world Street View imagery 🌍 Unlock a new perspective with Aerial and Satellite Insights, part of Earth AI, in BigQuery and Agent Platform to quickly extract detailed info about our changing planet 🛰️ Build custom apps with Aerial and Satellite models in Model Garden from Earth AI that easily identify specific objects or scenes Read the blog at the link above and sign up for early access.

Google Maps Platform

12,355 просмотров • 3 месяцев назад

Guys i think it's finally ready! You pick any area on a map and get a faithful, 1:1 3D version of it, in 30 seconds or less. It's tested, it works, and I'm shipping it this week once I'm done tightening the optimization. Right now it generates flawlessly anywhere in France and across the EU (the OSM data is just so rich here), and I'm adding US support as I type this. Honestly still a little amazed it works this well🙏 What should I call it? L'atelier? Bbox? Drop your suggestions below 👇
0:40

Guys i think it's finally ready! You pick any area on a map and get a faithful, 1:1 3D version of it, in 30 seconds or less. It's tested, it works, and I'm shipping it this week once I'm done tightening the optimization. Right now it generates flawlessly anywhere in France and across the EU (the OSM data is just so rich here), and I'm adding US support as I type this. Honestly still a little amazed it works this well🙏 What should I call it? L'atelier? Bbox? Drop your suggestions below 👇

Playda

31,052 просмотров • 1 месяц назад

testing new model on binance sol perp, didnt write a single line of code, all writen by AI (of course i fed it tons of my old code to it). the per-level toxicity avoid learned (ai learned) from wintermute hl wallet works insanely well agi is here
5:02

testing new model on binance sol perp, didnt write a single line of code, all writen by AI (of course i fed it tons of my old code to it). the per-level toxicity avoid learned (ai learned) from wintermute hl wallet works insanely well agi is here

Takopi🐙

67,477 просмотров • 6 дней назад

I’m testing the new 4-step Turbo LoRA for MiniMax. With just 16GB of VRAM, I can generate 15 seconds of 480p video in around 1 minute and 30 seconds! I also noticed something pretty funny if you use a song as an audio reference and tell the character to dance to it, the character actually follows the rhythm of the music surprisingly well.
0:30

I’m testing the new 4-step Turbo LoRA for MiniMax. With just 16GB of VRAM, I can generate 15 seconds of 480p video in around 1 minute and 30 seconds! I also noticed something pretty funny if you use a song as an audio reference and tell the character to dance to it, the character actually follows the rhythm of the music surprisingly well.

Alex

49,428 просмотров • 10 дней назад

*Warning* the video is somewhat NSFW, I'm testing another new technique that can generate a Gaussian Splatting model using only three viewing angles! It already works, but there's still a kind of ghosting/doubling effect that I need to fix.
1:29

*Warning* the video is somewhat NSFW, I'm testing another new technique that can generate a Gaussian Splatting model using only three viewing angles! It already works, but there's still a kind of ghosting/doubling effect that I need to fix.

Alex

13,425 просмотров • 3 месяцев назад

i'm biased but Always Sunny. every element works so well within the context of the show. "did you fuck my mom?" mac's 'philly tradition'. the iconic scene of danny devito and the couch.their childhood backgrounds.mac and charlie throwing rocks on the train
0:35

i'm biased but Always Sunny. every element works so well within the context of the show. "did you fuck my mom?" mac's 'philly tradition'. the iconic scene of danny devito and the couch.their childhood backgrounds.mac and charlie throwing rocks on the train

female dennis reynolds

259,770 просмотров • 1 год назад

Staking testing is in works. Excited to get this in hands of public so you can test it out too. Devs are working round the clock and i mean it literally.
0:43

Staking testing is in works. Excited to get this in hands of public so you can test it out too. Devs are working round the clock and i mean it literally.

Zack | Pumpkin 🎃

13,845 просмотров • 1 год назад

This is one of the creepiest things on Google Earth... They say it's caused by satellite imagery being mirrored to fill in the gaps. Nevertheless... creepy
0:49

This is one of the creepiest things on Google Earth... They say it's caused by satellite imagery being mirrored to fill in the gaps. Nevertheless... creepy

Jay Anderson

260,555 просмотров • 1 месяц назад

can i interest you in the unreleased footage of the darkroom scene?
1:00

can i interest you in the unreleased footage of the darkroom scene?

mar.

39,963 просмотров • 11 месяцев назад

every AI model breaks on mirrors.. so i built the hardest possible test: a room made entirely of mirrors, a camera, and 30 unbroken seconds of movement.. Seedance 2.5 didn't blink.. prompt below
0:30

every AI model breaks on mirrors.. so i built the hardest possible test: a room made entirely of mirrors, a camera, and 30 unbroken seconds of movement.. Seedance 2.5 didn't blink.. prompt below

Abdul Șhakoor

10,852 просмотров • 9 дней назад

I think Noah Kahan singing this is me trying by Taylor Swift works so well because it truly feels like one of his own songs. The lyrics. The melody. It all works so so well in his voice. His version of it is stunning
1:28

I think Noah Kahan singing this is me trying by Taylor Swift works so well because it truly feels like one of his own songs. The lyrics. The melody. It all works so so well in his voice. His version of it is stunning

Noah Kahan Archive

117,701 просмотров • 5 дней назад

Unbeatable! Savings that gives you a 50% BOOST on what you put in! Can you get it? Just one of scores of ways to boost your interest in the new episode of the Martin Lewis Podcast. Listen to it all on (+Apple, Spotify & all the usual pod players)
1:10

Unbeatable! Savings that gives you a 50% BOOST on what you put in! Can you get it? Just one of scores of ways to boost your interest in the new episode of the Martin Lewis Podcast. Listen to it all on (+Apple, Spotify & all the usual pod players)

Martin Lewis

367,941 просмотров • 9 месяцев назад

still experimenting with LoRA based on the Thinking Machines configuration and just implemented it in colab. In this notebook I set up a fine tune of Qwen/Qwen3-0.6B on the OpenR1-Math dataset with lora rank of 1. with this setup you can get the same reward accuracy as full fine-tuning, at a fraction of vram usage.
0:45

still experimenting with LoRA based on the Thinking Machines configuration and just implemented it in colab. In this notebook I set up a fine tune of Qwen/Qwen3-0.6B on the OpenR1-Math dataset with lora rank of 1. with this setup you can get the same reward accuracy as full fine-tuning, at a fraction of vram usage.

Ben Burtenshaw

25,624 просмотров • 10 месяцев назад

-Wiggle physics enabled -There is a crazy wind in the scene that constantly changes its value -Tomas's head is animated -4 collision objects of different shapes, and all of this works in real-time. The hybrid XPBD solver works much more stably than the old one, I already really want to release the add-on, but I want to make sure it is ready for production cause you know, I'm a heavy overthinker lol
0:30

-Wiggle physics enabled -There is a crazy wind in the scene that constantly changes its value -Tomas's head is animated -4 collision objects of different shapes, and all of this works in real-time. The hybrid XPBD solver works much more stably than the old one, I already really want to release the add-on, but I want to make sure it is ready for production cause you know, I'm a heavy overthinker lol

foxtomas

28,402 просмотров • 2 месяцев назад

I'm testing my AI render projection shader on de_dust2 from CS 1.6. I updated the shader again so it works well with large open spaces, not just interiors.
1:10

I'm testing my AI render projection shader on de_dust2 from CS 1.6. I updated the shader again so it works well with large open spaces, not just interiors.

Alex

28,604 просмотров • 6 месяцев назад

This is Segment Anything for #GaussianSplatting running on the web... but it comes with a caveat. This is SplatEdit, a powerful #3DGS editing software built on top of SuperSplat. Last week, it added a segmentation selection feature powered by Segment Anything, running locally in the browser. Just upload a splat, click on the "mug" icon and select entire objects in the scene. The problem is that everything after the object is also selected (as you can see in the video). To fix that, just use the brush tool to subtract the unwanted portion of the scene (hold CTRL while brushing). I feel we are getting closer and closer to having more editing control on these amazing fuzzy clouds 😉
0:37

This is Segment Anything for #GaussianSplatting running on the web... but it comes with a caveat. This is SplatEdit, a powerful #3DGS editing software built on top of SuperSplat. Last week, it added a segmentation selection feature powered by Segment Anything, running locally in the browser. Just upload a splat, click on the "mug" icon and select entire objects in the scene. The problem is that everything after the object is also selected (as you can see in the video). To fix that, just use the brush tool to subtract the unwanted portion of the scene (hold CTRL while brushing). I feel we are getting closer and closer to having more editing control on these amazing fuzzy clouds 😉

Gabriele Romagnoli

14,634 просмотров • 1 месяц назад

I was given the chance to animate and model for Hakos Baelz's trip to NinjaTokyo! 🎲🐀 #BaelzBrush Here's a little segment of it! You can watch the full version on her youtube channel! (link below)
1:01

I was given the chance to animate and model for Hakos Baelz's trip to NinjaTokyo! 🎲🐀 #BaelzBrush Here's a little segment of it! You can watch the full version on her youtube channel! (link below)

Yoolie

25,549 просмотров • 1 год назад