💃 Dancing in the wild with StyleGAN 💃

👉StyleGAN-based animations for AR/VR apps

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Video based motion retargeting
✅A StyleGAN architecture based
✅Novel explicit motion representation
✅SOTA qualitatively & quantitatively

More: https://bit.ly/3CZbL1W

🪀TensoRF: the 4D evolution of NeRF 🪀

👉TensoRF, a novel radiance fields via 4D-tensor: 3D voxel grid with per-voxel multi-channel feats.

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅VM decomposition technique
✅Low-rank tensor factorization
✅Lower memory footprint (speed)
✅TensoRF is the new SOTA in R.F.
✅Code under the MIT License

More: https://bit.ly/3qffZgI

🔼 GAN-meshes without key-points 🔼

👉ETH unveils a GAN framework for generating textured triangle meshes without annotations

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Generative of textured meshes
✅3D generator for all categories
✅3D pose estimation framework
✅Code licensed under MIT License

More: https://bit.ly/3qfH9nJ

🐯 S.S. Latent Image Animator 🐯

👉Self-supervised autoencoder to animate unseen images by linear navigation in latent

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Latent Image Animator
✅Linear displacement in latent
✅SOTA: VoxCeleb, Taichi, TED-talk
✅Source code (soon) available

More: https://bit.ly/36pgLAC

🪨 Google URF for neural-synthesis 🪨

👉Sequence of RGB + Lidar -> 3D surfaces and novel RGB images synthesized

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Extending Neural Radiance Fields
✅Leveraging asynch. lidar data
✅Addressing exposure variation
✅Leveraging segmentations for sky
✅SOTA #3D reconstructions/synthesizes

More: https://bit.ly/3L2vTDb

🚛 AV2: next-gen. self driving 🚛

👉One of the biggest dataset ever for #autonomousdriving

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅1k seq. of multimodal data
✅3D annotations, 26 categories
✅20k lidar & map-aligned pose
✅250k challenging interactions
✅HD Map: 3D lane & crosswalk
✅CC BY-NC-SA 4.0 license

More: https://bit.ly/3trx3lw

🤖CaTGrasp in Clutter from Simulation🤖

👉Task-relevant grasping: trained solely in simulation with synthetic + SS. hand-object interaction

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Novel cat-level, relevant grasping
✅S.S. hand-object-contact
✅Tiny objects from dense clutter
✅Train-simulation -> to real
✅Source code under Apache 2.0

More: https://bit.ly/3L2YVCo

🛼 Drive & Segment without Supervision 🛼

👉Learning pixel-wise semantic seg. on non-curated data collection by cars (cameras + LiDAR) driving around a city

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Cross-modal unsupervised
✅Synchronized LiDAR & RGB
✅Object proposal on LiDAR points
✅SOTA, significant improvements

More: https://bit.ly/3L0wWTW

🌍 NeRF-free Neural Rendering 🌍

👉A simple 2D-only method with a single pass of a neural network

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Synthesis with NO 3D reasoning
✅Autoregressive & masked transf.
✅Pose -> object, object -> pose
✅Attention: branching attention
✅Source code under MIT License

More: https://bit.ly/3JC7unt

🤓👌Hey, TAKE OFF my eyeglasses! 😙👌

👉A novel framework to remove eyeglasses as well as their cast shadows from faces

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Novel mask-guided multi-step network
✅Leveraging 3D synthetic data only
✅Synthetic portraits with supervisions
✅Eyeglasses & shadows simultaneously

More: https://bit.ly/3IvQzlf

🏥 #AI models/dataset for open surgery 🏥

👉Multi-task #AI model/dataset of real-time surgical behaviors, hands, and tools.

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Annotated Videos Open Surgery
✅Largest dataset of open surgical
✅2k clips and 23 procedures
✅12k annotations, 11k+ keypoints
✅Models/Dataset soon available!

More: https://bit.ly/3tvDdkK

🥽 #metaverse in 1991 🥽

👉Q: is #VR the technology that developed least in the last 30 years? 🤔

Discussion: https://bit.ly/3txWF07

🫕NeRFusion: Large-Scale Reconstruction🫕

👉Efficient large-scale reconstruction & photo-realistic rendering

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Frame-by-frame R.F.
✅Neural reconstruction
✅Real-time at 20+ fps
✅SOTA on indoor / objects

More: https://bit.ly/3iyfoCo

☕ORViT for understanding tasks☕

👉ORViT: object-centric approach that extends ViT layers incorporating object representations

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Spatio-temporal through the net
✅''Object-Region Attention''
✅''Object-Dynamics" module
✅Code just released! Apache 2.0

More: https://bit.ly/3wAUavW

🪅Insane Neural Sketching from #MIT🪅

👉Line drawing generation as unsupervised image translation with various losses

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Unpaired method for line drawing
✅Geometry loss to predict depth
✅Semantic loss to match CLIP feats
✅SOTA on unpaired translation/generation
✅Code and Models under MIT License

More: https://bit.ly/36JRr8A

🏔️MPS-Net: new SOTA for #3D human🏔️

👉MPS-Net: accurate & temporally coherent 3D human pose/shape from video

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅MoCA: visual cues from motion
✅HAFI to mix past/future feats
✅Stronger temporal correlation
✅SOTA on multiple datasets

More: https://bit.ly/3uAI5EB

🤿Transfiner: hyper-detailed segmentation🤿

👉Mask Transfiner: #AI for HQ & efficient instance segmentation

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅Transfiner: HQ segmentation
✅HQ seg. via quadtree structure
✅SOTA & extreme details
✅Code under MIT License

More: https://bit.ly/3KVzseM

🥙 DualStyleGAN: SOTA in style transfer🥙

👉Flexible control of dual styles of face domain and extended artistic portrait domain

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅High-resolution (1024*1024)
✅Intrinsic/extrinsic style path
✅Hierarchical style manipulation
✅Novel progressive fine-tuning
✅Source code under MIT License

More: https://bit.ly/3uS26Xp

🍚 GTR: Global Tracking Transformers 🍚

👉UTexas + Apple: transformer for global multi-object tracking

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅GTR operates on any object
✅Few frames->global trajectories
✅SOTA on detectors for any object
✅Code under Apache License 2.0

More: https://bit.ly/3DiqkxF

🧠E2E Perception for #selfdrivingcars🧠

👉HybridNets: multi-task net with several key optimizations

𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬:
✅End-to-end perception network
✅Traffic, lane, object detection
✅Drivable segmentation area
✅Real-time on embedded systems
✅Source code under MIT License

More: https://bit.ly/3JMk8Az