Alexander Richard

github repository: github.com/facebookresear…

‼️500h of 3D motion data released‼️ Our team at the Codec Avatars Lab just released a large scale dataset of 3D tracked human motion, including audio and text annotations. Check it out here: meta.com/emerging-tech/…

Join our team for an internship in 2025!

Working on your next big speech paper but still looking for a suitable dataset? Check out EARS: 100h of full-band expressive, anechoic recordings of speech from 107 speakers with 22 different emotions, 7 different reading styles, and more. sp-uhh.github.io/ears_dataset/

If you're at #CVPR2024 this week, don't miss talking to @AlexRichardCS we are hiring for postdoctoral and full time positions! metacareers.com/jobs/380024285… metacareers.com/jobs/803309161…

Want to make computers that can see *and* hear? Come to the Sight and Sound workshop on Monday, 6/17 at #CVPR2024! Schedule: sightsound.org Invited talks by: @HildeKuehne, @ShyamGollakota, @RuohanGao1, @TengdaHan, @SamuelPClarke, @AlexRichardCS

📢Curious about the future of 3D scene understanding? Join us at the 1st Workshop on Multimodalities for 3D Scenes @CVPR! Learn about the latest research on using vision, audio, touch, and language to understand 3D scenes around us. 📅June 17, 1:30-5:20PM multimodalitiesfor3dscenes.github.io

We have 4 amazing speakers (@xiaolonw, @andrewhowens, @KaterinaFragiad, Linda Smith) to cover different modalities, from vision, language, robotics, touch, and cognitive science. Co-organizing with Angel Chang, @_krishna_murthy, @AlexRichardCS, Kristen Grauman. #AI #CVPR

Real Acoustic Fields are here! Check out or dataset of densely captured room impulse responses paired with multi-view images! See you all on CVPR :) project page: facebookresearch.github.io/real-acoustic-… arxiv: arxiv.org/abs/2403.18821

Always good to see the progress we’re making in our labs in NYC and (in the case of these photos) Pittsburgh...

@angrypenguinPNG @_akhaliq Nice. As Mentioned in the google collab, lip sync gets significantly better if you have at least 10 seconds of audio :)

Quick test of Meta's Photoreal Embodiment 🤖 Audio to Synthesized Human Movement in under 10 mins ⚡️ Weights, Code, and Dataset: github.com/facebookresear…

@danveloper The restriction comes from our data collection: we only train on dyadic conversations, so the model has never seen 3+ people in the same environment.

@AlexRichardCS This is really cool, Alex! Is the reason it’s constrained to dyadic conversation because of speaker diarization problems across a long running conversation?

Motion generation for photorealistic avatars? Say no more! Check out how we animate full body avatars exclusively from audio input! Paper: arxiv.org/abs/2401.01885 Project page: people.eecs.berkeley.edu/~evonne_ng/pro… Dataset + Code: github.com/facebookresear…

3D body models now have sound! We demonstrate 3D spatial audio synthesis for 3D full body models. (NeurIPS 2023 Spotlight) Paper: openreview.net/pdf?id=zQTi3pz… Dataset + Code (to be released): github.com/facebookresear…

Work done in collaboration with @AlexRichardCS, @ovrdr, Vamsi Ithapu, @NataliaNeverova, Kristen Grauman, and Andrea Vedaldi @MetaAI @RealityLabs The poster session is Tues-PM and I will also give a talk at the sight and sound workshop Monday 4:30pm.

A new generalized and universal audio-visual speech enhancement model, powered by SSL. A single model for denoising, source sep, inpainting and lip-reading! Checkout the paper and demo video!! 🗣️🤖🔊 More details below with @mhnt1580

Diffused Heads: Diffusion Models Beat GANs on Talking-Face Generation abs: arxiv.org/abs/2301.03396 project page: mstypulkowski.github.io/diffusedheads/

📢Excited to share our recent #ECCV2022 paper: "LiP-Flow: Learning Inference-Time Priors for Codec Avatars via Normalizing Flows in Latent Space" with Shugao Ma, @akcalakcal, Stanislav Pidhorskyi, @AlexRichardCS, Shih-En Wei, Jason Saragih and @OHilliges.

Happy to be on the list of outstanding reviewers for @eccvconf with my dear friends and coworkers @aayushbansal and @c_richardt Congratulations!

**Dataset Release!** We released high-quality 3D face data of 13 persons captured with up to 150 cameras while performing 100+ facial expressions! github.com/facebookresear…