Matteo Bortolon

PhD Student in Computer Vision, at TeV @FBK_research & @PAVIS_IIT

[TeV - Technologies of Vision @FBK @TeV_FBK]

🚀 Introducing FreeGrasp: Free-form language-based robotic reasoning and grasping 🤖

🎥 Watch a UR5e robotic arm executing complex grasping tasks based on natural language commands!

🔗 Learn more: https://t.co/0q26CE6KuQ
🧵👇

[PAVIS IIT @PAVIS_IIT]

Paper and code already available, please check...

Project page: https://t.co/F8nHsXcJDK
arXiv: https://t.co/0rTKVGhkiB

Kudos to the authors: @alessiodelbue @Poiex @sttujames @mbortolon97 and Theo!

2/2

[PAVIS IIT @PAVIS_IIT]

We present 6DGS, our new @eccvconf paper on camera pose estimation using an NVS model! 📸🔬Our approach proposes a novel "radiant Ellicell" for 3DGS to quickly compare the image with the model.

Result? Faster, more accurate one-shot pose recovery.

🧵1/2

[PAVIS IIT @PAVIS_IIT]

Our new @ieee_ras_icra 2024 shows we can estimate the camera pose from a #NeRF model at 34fps from a single image 🚀🚀🚀 and without the need for a rotation and translation for initialization! 💡💡💡

🎬This short video shows how we did it... 🧵