PulseAugur / Brief
EN
LIVE 03:36:19

Brief

last 24h
[4/4] 224 sources

Multi-source AI news clustered, deduplicated, and scored 0–100 across authority, cluster strength, headline signal, and time decay.

  1. TOOL · Mastodon — mastodon.social English(EN) ·

    🔥🚀 Introducing # TorchCodec 0.14, where you can now decode # HDR # video using your # toaster oven's CPU or a NASA-grade # CUDA , while a "fast" WAV decoder spe

    TorchCodec has released version 0.14, introducing the capability to decode High Dynamic Range (HDR) video using a wide range of hardware, from standard CPUs to high-performance CUDA GPUs. This update also includes a faster WAV decoder, potentially improving performance for audio tasks. AI

    IMPACT Enhances video processing capabilities with broader hardware support, potentially impacting media production and AI-driven video analysis tools.

  2. FRONTIER RELEASE · X — Luma Labs (video gen) English(EN) · · [5 sources]

    Reframe handles the backdrop. V2V handles the SKU. HDR generation and EXR export handle the mood. One brand system, every market, every format, generated consis

    Luma Labs has released updates to its video generation model, introducing enhanced features for professional use. The new capabilities include 1080p output, up to 20-second video generation with their V2V (video-to-video) model, and native HDR generation with EXR export for seamless integration into existing workflows. The updates also bring multi-keyframe control for precise direction of changes within a clip and improved motion transfer for realistic choreography. AI

    IMPACT Enhances professional video production workflows with advanced AI-driven generation and editing tools.

  3. RESEARCH · Hugging Face Daily Papers English(EN) · · [3 sources]

    ExpoCM: Exposure-Aware One-Step Generative Single-Image HDR Reconstruction

    Researchers have developed ExpoCM, a new framework for reconstructing high dynamic range (HDR) images from single low dynamic range inputs. This method addresses the challenges of detail loss in over-exposed and noise in under-exposed areas by reformulating the problem as a Probability Flow ODE. ExpoCM uses exposure-aware consistency trajectories and an exposure-guided loss function to improve image quality and significantly speed up inference times compared to existing diffusion models. AI

    ExpoCM: Exposure-Aware One-Step Generative Single-Image HDR Reconstruction

    IMPACT Offers a faster and more accurate method for HDR image reconstruction, potentially improving visual quality in photography and computer vision applications.

  4. RESEARCH · arXiv cs.CV English(EN) · · [3 sources]

    FDIM: A Feature-distance-based Generic Video Quality Metric for Versatile Codecs

    Researchers have introduced FDIM, a new video quality metric designed to evaluate both traditional and neural video codecs across Standard Dynamic Range (SDR) and High Dynamic Range (HDR) content. FDIM utilizes a hybrid approach, combining deep learning for multi-scale feature extraction with hand-crafted features to capture a wide range of distortions. Trained on a large dataset of over 16,000 video sequences, FDIM has demonstrated strong generalization capabilities across various codecs and content types, correlating well with subjective quality assessments. AI

    FDIM: A Feature-distance-based Generic Video Quality Metric for Versatile Codecs

    IMPACT Provides a generalized metric for evaluating neural video codecs, potentially accelerating research and development in video compression.