Abdullah Hamdi
@Eng_Hemdi • 10,272 subscribers
research scientist @KAUST_News 3D-AI+agents | prev: postdoc @Oxford_VGG 🇬🇧 | @fihmai founder | my @tedX talk about AI inequality: https://t.co/Y24DtOs33b
Videos
having worked on computer vision for +10 years , I have been asking myself lately how to solve EVERY computer vision problem with whatever LLM tech we have now. First step: image segmentation. meet LLMSeg, segmenting images with LLMs and B-splines loops. Basically no need for trained segmentation heads that do not generalize well, just use the MLLMs superior generalization to perform segmentation directly with no tools or CLI , just a simple pre-scripted forward loop with refinement process based on overlays. B-splines loops for base masks , and B-splines loops for positive and negative masks refinment ----> THEY ACTUALLY CONVERGE to more accurate masks 21 MLLMs tested , better "vision+reasoning" means better segmentation . any "decent" MLLM can do segmentation it turns out. website: paper: code: sounds on 🔊
Abdullah Hamdi21,036 görüntüleme • 6 gün önce
📢Happy to present Convex Splatting, a novel way for 3D reconstruction based on 3D smooth convexes. For the first time, a splatting-based method reaches the quality of NeRF sota methods but with real-time rendering and few primitives!! I expect this to replace Gaussian splatting for 3D in the coming months. CODE RELEASED TODAY! joint work with collaborators from Université de Liège Visual Geometry Group (VGG) , KAUST Computer Vision Lab (IVUL) a thread 🧵 1/n
Abdullah Hamdi71,661 görüntüleme • 1 yıl önce
Daha fazla içerik yok.