HDR
Datasets
All datasets matching “HDR”S2R-HDR🌐Homepage | 📖Arxiv | GitHub
✨Dataset Summary
S2R-HDR is a large-scale synthetic dataset for high dynamic range (HDR) reconstruction tasks. It contains 1,000 motion sequences, each comprising 24 images at 1920×1080 resolution, with a total of 24,000 images. To support flexible data augmentation, all images are stored in EXR format with linear HDR values. The dataset is rendered using Unreal Engine 5 and our custom pipeline built upon XRFeitoria, encompassing diverse dynamic… See the full description on the dataset page: https://huggingface.co/datasets/iimmortall/S2R-HDR.S2R-HDR-2🌐Homepage | 📖Arxiv | GitHub
✨Dataset Summary
S2R-HDR is a large-scale synthetic dataset for high dynamic range (HDR) reconstruction tasks. It contains 1,000 motion sequences, each comprising 24 images at 1920×1080 resolution, with a total of 24,000 images. To support flexible data augmentation, all images are stored in EXR format with linear HDR values. The dataset is rendered using Unreal Engine 5 and our custom pipeline built upon XRFeitoria, encompassing diverse dynamic… See the full description on the dataset page: https://huggingface.co/datasets/iimmortall/S2R-HDR-2.hdr-demo-clips
HDR Demo Clips (Lightricks SDR→HDR)
Paired SDR (input) / HDR (output) frame sequences from the Lightricks SDR-to-HDR pipeline (IC-LoRA on LTX-2).
Each clip contains:
hdr_exr/frame_XXXXX.exr — HDR output (f16, linear Rec.709/sRGB primaries, scene-referred)
sdr_png/frame_XXXXX.png — SDR input (8-bit sRGB, display-referred)
thumbnail.jpg — 280px preview from the middle frame
Dimensions: HDR is symmetrically cropped from SDR to match model-friendly dimensions (typically 28–56px… See the full description on the dataset page: https://huggingface.co/datasets/oumoumad/hdr-demo-clips.Skydome_HDRIDownload the waymo_skydome folder and put it in data.
HDRTV1K
HDRTV1K
HDRTV1K is a dataset for SDRTV-to-HDRTV conversion and related HDR imaging research, hosted on Hugging Face:
Hugging Face Dataset: chxy95/HDRTV1K
This dataset is associated with the HDRTVNet / HDRTVNet++ project line for SDRTV-to-HDRTV conversion. It contains HDR and SDR image data for training and testing, and is suitable for tasks such as HDR reconstruction, SDR-to-HDR translation, inverse tone mapping, and other low-level vision applications. According to the Hugging… See the full description on the dataset page: https://huggingface.co/datasets/chxy95/HDRTV1K.HDR_Pretrain
HDR-Pretrain: A Pretraining Dataset for 3D HDR Reconstruction
Paper | Code | Project page
HDR-Pretrain is a synthetic multi-view dataset for pretraining 3D high dynamic range (HDR) reconstruction models. It provides multi-exposure low dynamic range (LDR) images and corresponding HDR references to support learning 3D HDR scene representations from multi-view observations and rendering HDR images from novel viewpoints.
The dataset contains 168 Blender-rendered indoor scenes. Each… See the full description on the dataset page: https://huggingface.co/datasets/Bugjudger/HDR_Pretrain.