AI Briefing#computer-vision
The latest AI and developer news about #computer-vision, with the original source and a short summary.
The latest AI and developer news about #computer-vision, with the original source and a short summary.
Mistral AI has released OCR 4.1, featuring paragraph-level Bounding Box extraction and structural block labeling capabilities.
Released worldproof, an open-source tool that diagnoses prediction errors in world models and measures the discriminative limitations of pixel-based metrics.
The 7B-scale open model SenseNova-Vision has been released, capable of performing various vision tasks without separate heads.
WorldClaw generates editable 3D open worlds from a single prompt.
GaussianSplatting.jl 2.0 has been released, offering multi-GPU support and enhanced geometric precision.
Multimodal models developed more effectively when modalities were trained together from the beginning.
F.R.A.N.T.Z. extracts scene semantics from monocular video without synthetic data or task-specific training.
The study comparatively analyzed preference alignment methods and datasets for multimodal LLMs and proposed BDHS.
Parallel Decoding Distillation predicts multiple denoising steps at once to improve the speed and diversity of video generation.
Microsoft has unveiled Mage, a family of efficient multimodal models built at a 4B parameter scale.
The AI tool Starbucks introduced to automate inventory was scrapped after 9 months due to technical flaws.
FeyNoBg, an open-source background removal model and training library achieving SOTA performance, has been released.
NVIDIA has unveiled Cosmos-H-Dreams, a real-time generative simulator that accelerates the training and evaluation of surgical robots.
Nvidia has announced SANA-Video 2.0, which accelerates high-resolution long-form video generation through hybrid attention technology.
GOT-OCR2.0, an end-to-end OCR model that processes complex visual information such as formulas, tables, and sheet music with a single model, has been released.
It is an AI agent tool that analyzes a single image and reconstructs it into 3D model code based on Three.js.
An 8B-parameter infographic generation model with enhanced text and layout editing capabilities has been released.
Xiaomi has unveiled a robot foundation model trained on 100,000 hours of embodiment-free data.
We propose RayRoPE, a positional encoding scheme with SE(3) invariance and geometric adaptability for multi-view transformers.
A multi-contrast MRI reconstruction framework that works without k-space data by modeling content and style has been released.
LG AI Research has developed SISE, a new image recognition explanation technology that overcomes the limitations of existing XAI models.
LG AI Research used Contrastive Learning to build an unsupervised anomaly detection model with higher accuracy than supervised learning.
LG AI Research has revealed the technical principles of EXAONE, a multimodal model that understands text and images bidirectionally.
Introduces the latest trends in multimodal representation learning and NeRF technology that drew attention at CVPR 2022.