Intelligent eyewear is coming this fall! ✨
It was a privilege to be onstage at #GoogleIO to show our new audio glasses with #Gemini and share a first look at two designs from the upcoming collections.
I’m incredibly proud of our team and the partnership with @SamsungMobile and iconic eyewear brands @WarbyParker and @_GentleMonster_ to bring hands-free help to your everyday life.
Read more: https://t.co/CJolDrVu7P
Watch I/O Keynote: https://t.co/gCPsGvcVwn
Meet D4RT 🎯
A unified model for next-gen scene reconstruction and tracking.
🧠 The Secret Sauce: Incredible results thanks in part to diverse, realistic synthetic data from our team. The results are wild. 🤯
We're helping AI to see the 3D world in motion as humans do. 🌐
Enter D4RT: a unified model that turns video into 4D representations faster than previous methods - enabling it to understand space and time. This is how it works 🧵
#SyntheticData powers cutting-edge algorithms on #AndroidXR like Hand Tracking, Face Tracking, and more. 🦾
At #CVPR2025 ? Attend our tutorial (June 12th) to dive deep into Android XR's Perception stack.
Full schedule here: https://t.co/CbdyAMfQlj
#AndroidXR is officially here! 🎉 Thrilled to see the developer excitement and passion to build immersive experiences. Huge thanks to the @GoogleARVR team and our incredible partners @SamsungMobile and @Qualcomm and more. ➡️ https://t.co/0ljK0syK2F
It turns out that a synthetic-only face prior is great for novel view synthesis 🦾. If you're heading to SIGGRAPH Asia 🇯🇵, don't miss Cᴀꜰᴄᴀ. Great work, @mc_buehler!
A purely synthetic face prior can generalize to casual in-the-wild captures and stylized faces 🤯
I'm happy to share our @SIGGRAPHAsia paper “Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures.”
https://t.co/UPNMexPbRc
Coming to #CVPR23? Come to my talk on Sunday morning at the GAZE workshop https://t.co/6B0qJ0n4B8 to hear about Synthetic Data for Gaze Estimation & More. Find out what worked 🙌🏻, what didn't 🤔, and what's next 🚀.
At #ECCV2022? Come try our our face tracker, powered by #syntheticdata and some real-time model-fitting magic 🪄.
🖼️ Poster 59 in Thursday's morning session.
🔗 https://t.co/aoqg2BVrNS
Excited to share our work on #syntheticdata, generating dynamic expression-based wrinkles for digital faces, to appear at #WACV2023!
With @Friggog, @errollw, and Tadas Baltrušaitis @Microsoft Mixed Reality & AI Cambridge
📝https://t.co/3oSjCBedHK
🧵1/8
Very proud to share more about our work on 𝗗𝗲𝗻𝘀𝗲 𝗙𝗮𝗰𝗲 𝗟𝗮𝗻𝗱𝗺𝗮𝗿𝗸𝘀, which will appear at #ECCV2022. Check it out! 👇
🔗 https://t.co/aoqg2BVrNS
🎥 https://t.co/AL7fcDNPTv
📝 https://t.co/4fGrgxAYVF
@cvtalks That's a really cool dataset! The community will love it!
However, I'd argue that #syntheticdata can give you more consistent labels, and greater training data diversity.
Join me tomorrow at the SyntML workshop at #CVPR2022 to hear about how #syntheticdata helps us solve all sorts of face-related computer vision problems. These aren't toy problems, they are the real deal!
https://t.co/zkGeKSHMLB
@91amin91 Mediapipe is awesome, and I love their Attention Mesh design. But I think our synthetic training data is just better than their real data😉
We have nothing to share at this time unfortunately, but please stay tuned for further announcements🙂
I'm delighted to share our work on dense landmarks! https://t.co/4fGrgxAYVF
🌟 Dense landmarks are all you need for accurate 3D face reconstruction and performance capture
💪 No surprise: the secret sauce is synthetic data
🏃 It's fast, running at 150FPS on a single CPU thread