Video

229 public types in this namespace, organized by kind.

Models & Types (108)

Type Summary
ABME<T> ABME: asymmetric bilateral motion estimation for video frame interpolation (Park et al., ICCV 2021).
AMT<T> AMT: all-pairs multi-field transforms for efficient frame interpolation (Li et al., CVPR 2023).
AVID<T> AVID diffusion-based video inpainting supporting arbitrary-length videos.
BSVD<T> BSVD bidirectional streaming video denoising with real-time buffers.
BasicVSRPlusPlus<T> BasicVSR++ (Basic Video Super-Resolution++) for temporal video super-resolution.
BasicVSR<T> BasicVSR: baseline bidirectional recurrent video super-resolution.
BiMVFI<T> BiM-VFI: bidirectional motion field-guided frame interpolation for video with non-uniform motions (Seo, Oh and Kim, CVPR 2025).
Cutie<T> Cutie: Cutting-edge Video Instance Segmentation with transformer memory.
DAMVSR<T> DAM-VSR: video super-resolution on Stable Video Diffusion that takes appearance from an image super-resolution model and motion from the low-resolution video.
DIFRINT<T> DIFRINT: Deep Iterative Frame Interpolation for Full-frame Video Stabilization.
DKM<T> DKM dense kernelized feature matching for geometry estimation.
DOVE<T> DOVE: efficient one-step diffusion for real-world video super-resolution, fine-tuned from CogVideoX1.5-5B.
DPFlow<T> DPFlow dual-pyramid framework combining image and feature pyramid advantages.
DUT<T> DUT deep unsupervised trajectory-based video stabilization.
DepthAnythingV2<T> Depth Anything V2 for monocular depth estimation.
DualXVSR<T> DualX-VSR: a dual axial spatial x temporal transformer for real-world video super-resolution without motion compensation.
DynamiCrafter<T> DynamiCrafter: animating open-domain images with video diffusion priors.
E2FGVI<T> E2FGVI - End-to-End Framework for Flow-Guided Video Inpainting.
EDVR<T> EDVR: Video Restoration with Enhanced Deformable Convolutional Networks.
EMAVFI<T> EMA-VFI: extracting motion and appearance via inter-frame attention for efficient video frame interpolation (Zhang et al., CVPR 2023).
EnhancementProgress Reports progress during video enhancement operations.
FILM<T> FILM (Frame Interpolation for Large Motion) model for high-quality frame interpolation.
FLAVR<T> FLAVR: flow-agnostic video representations for fast frame interpolation (Kalluri et al., WACV 2023).
FastDVDNet<T> FastDVDNet: Towards Real-Time Deep Video Denoising Without Flow Estimation.
FiganMultiScaleFlow<T> FIGAN’s multi-scale residual flow estimator and its occlusion-aware bidirectional synthesis.
Figan<T> FIGAN: frame interpolation with multi-scale deep loss functions and a generative adversarial network.
FlashVSR<T> FlashVSR: real-time diffusion-based streaming video super-resolution, distilled to one step from Wan 2.1.
FloRNN<T> FloRNN: unidirectional video denoising with a look-ahead recurrent module that mimics a backward one (Li et al., ECCV 2022).
FlowDiffuser<T> FlowDiffuser diffusion-based optical flow with iterative refinement.
FlowFormerPlusPlus<T> FlowFormer++ masked cost volume autoencoding with tile-based high-resolution flow.
FlowFormer<T> FlowFormer: A Transformer Architecture for Optical Flow.
FlowLens<T> FlowLens optical-flow-guided video inpainting with flow completion.
FuSta<T> FuSta hybrid full-frame video stabilization with warping and outpainting.
FuseFormer<T> FuseFormer transformer-based video inpainting with fine-grained spatial-temporal fusion.
GIMMVFI<T> GIMM-VFI: generalizable implicit motion modeling for video frame interpolation.
GMFlow<T> GMFlow (Global Matching Flow) for accurate optical flow estimation.
GaVS<T> GaVS gaze-aware video stabilization with saliency-weighted motion smoothing.
GroupedSpatialTemporalShift<T> The grouped spatial-temporal shift module: Shift-Net’s alignment mechanism, which replaces optical flow, deformable convolution and cross-frame attention with pure tensor shifting.
IART<T> IART: a recurrent video super-resolution transformer whose frame alignment is learned implicit resampling.
IFRNet<T> IFRNet: intermediate feature refine network for efficient frame interpolation (Kong et al., CVPR 2022).
IQVFIDistillation<T> IQ-VFI’s knowledge-distillation objective (Hu, Jiang, Zhong, Wang and Zheng, CVPR 2024): a pyramid reconstruction loss plus acceleration and motion distillation from a privileged teacher.
IQVFI<T> IQ-VFI: image quality-aware video frame interpolation with degradation adaptation.
IconVSR<T> IconVSR: information-aggregation with coupled propagation for video super-resolution.
ImplicitQuadraticMotion<T> IQ-VFI’s implicit quadratic motion model (Hu, Jiang, Zhong, Wang and Zheng, CVPR 2024): modulates linear intermediate flows into quadratic ones using a latent acceleration prior.
InternVideo2<T> InternVideo2: Scaling Video Foundation Models for Multimodal Video Understanding.
InterpAnyClearer<T> InterpAny-Clearer: distance indexing and iterative reference-based estimation for clearer anytime video frame interpolation (Zhong et al., ECCV 2024), on its release’s DR-RIFE.
LiteDVDNet<T> LiteDVDNet high-speed video denoising (Ilchenko & Stirenko, IJIGSP 2025).
M2M<T> M2M: many-to-many splatting for efficient video frame interpolation (Hu, Niklaus, Sclaroff and Saenko, CVPR 2022).
MGLDVSR<T> MGLD-VSR: motion-guided latent diffusion for temporally consistent real-world video super-resolution.
MIAVSR<T> MIA-VSR: masked inter and intra-frame attention for efficient video super-resolution.
McnetDecomposition<T> MCnet’s motion/content decomposition: the asymmetric split that gives the model its name.
Mcnet<T> MCnet: future-frame prediction by decomposing a video into separate motion and content pathways.
MemFlow<T> MemFlow optical flow with memory for real-time historical motion aggregation.
MiDaS<T> MiDaS: Towards Robust Monocular Depth Estimation.
MoG<T> MoG: motion-aware generative frame interpolation combining flow and diffusion.
MoMo<T> MoMo: momentum diffusion model for bi-directional optical flow in frame interpolation.
NeuFlowV2<T> NeuFlow v2 high-efficiency optical flow on edge devices.
PSRT<T> PSRT-recurrent: a recurrent video super-resolution transformer that aligns frames by moving whole patches.
PWStableNet<T> PWStableNet pixel-wise warping video stabilization with per-pixel flow fields.
PerVFI<T> PerVFI: perception-oriented video frame interpolation.
ProPainter<T> ProPainter for video inpainting - removes unwanted objects and fills regions in video.
RAFT<T> Recurrent All-pairs Field Transforms (RAFT) for optical flow estimation.
RAPIDFlow<T> RAPIDFlow — Recurrent Adaptable Pyramids with Iterative Decoding for efficient optical-flow estimation (Morimitsu 2025, https://arxiv.org/abs/2501.04350).
RIFE<T> Real-time Intermediate Flow Estimation (RIFE) for video frame interpolation.
RPKNet<T> RPKNet recurrent partial kernel network with separable large kernels for flow.
RVM<T> RVM: Robust Video Matting for real-time human segmentation.
RVRT<T> RVRT: recurrent video restoration transformer with guided deformable attention.
RealBasicVSRSharp<T> RealBasicVSR-Sharp: the adversarial training stage of RealBasicVSR, trading PSNR for perceptual quality.
RealBasicVSR<T> RealBasicVSR: real-world video super-resolution with an image-cleaning module.
RealESRGANVideo<T> Real-ESRGAN Video: Real-ESRGAN real-world super-resolution applied to every frame of a video.
RealESRGAN<T> Real-ESRGAN (Real Enhanced Super-Resolution GAN) for image and video super-resolution.
RealViformer<T> RealViformer: a recurrent transformer for real-world video super-resolution built on channel attention.
RealisVSR<T> RealisVSR: detail-enhanced diffusion for real-world 4K video super-resolution, a Consistency-Preserved ControlNet on Wan 2.1.
RoMa<T> RoMa robust dense feature matching with DINOv2 foundation for pixel-dense warps.
SAM2<T> SAM 2 (Segment Anything Model 2) for promptable segmentation of images and videos.
SEARAFT<T> SEA-RAFT simple efficient accurate RAFT with mixture of Laplace loss.
SKFlow<T> SKFlow selective kernel attention for efficient optical flow estimation.
STMFNet<T> STMFNet: spatio-temporal multi-flow network for video frame interpolation.
STTN<T> STTN spatial-temporal transformer network for video inpainting with multi-scale attention.
SeedVR<T> SeedVR: seeding infinity in diffusion transformer towards generic video restoration, as the authors released it.
ShiftNet<T> Shift-Net: video denoising with grouped spatial-temporal shift (Li et al., CVPR 2023).
SlowFast<T> SlowFast Networks for Video Recognition.
SoftSplat<T> SoftSplat: softmax splatting for video frame interpolation (Niklaus and Liu, CVPR 2020).
StabStitch<T> StabStitch joint video stabilization and stitching for panoramic output.
StableVideoSR<T> StableVideoSR: stable diffusion with temporal conditioning for video super-resolution.
StreamDiffVSR<T> Stream-DiffVSR: low-latency streamable video super-resolution by auto-regressive diffusion.
SwinVFI<T> SwinVFI: Swin Transformer-based video frame interpolation.
TLBVFI<T> TLBVFI: token-level bidirectional video frame interpolation.
TTVSR<T> TTVSR: learning trajectory-aware transformer for long-range video super-resolution.
ThreeDMF<T> 3DMF 3D motion field video stabilization with depth-aware camera compensation.
TimeSformer<T> TimeSformer: Is Space-Time Attention All You Need for Video Understanding?
ToonCrafter<T> ToonCrafter generative cartoon interpolation for large non-linear animated motion.
TrainingProgress Reports training progress for video AI models.
UDVD<T> UDVD: unsupervised deep video denoising with a blind-spot network (Sheth et al., ICCV 2021).
UFM<T> UFM unified flow and matching demonstrating unified training beats specialized models.
UPRNet<T> UPR-Net: Unified Pyramid Recurrent Network for video frame interpolation.
UniMatch<T> UniMatch unified flow, stereo, and depth estimation with cross-task transfer.
Upscale4KAgent<T> 4KAgent (Zuo et al.
VFIMamba<T> VFIMamba state-space model for video frame interpolation with linear-complexity attention.
VFIT<T> VFIT: the video frame interpolation transformer (Shi, Xu, Liu, Chen and Yang, CVPR 2022).
VFIformer<T> VFIformer: video frame interpolation with a transformer (Lu, Cui, Tian, Jiang and Jia, CVPR 2022).
VRT<T> VRT: A Video Restoration Transformer for video super-resolution, deblurring, and denoising.
VideoCLIP<T> VideoCLIP model for video-text understanding and retrieval.
VideoFlow<T> VideoFlow multi-frame temporal cues for optical flow estimation.
VideoGigaGAN<T> VideoGigaGAN: towards detail-rich video super-resolution with large-scale GAN.
VideoMAE<T> Video Masked Autoencoder (VideoMAE) for video understanding and action recognition.
XMem<T> XMem: Long-Term Video Object Segmentation with Atkinson-Shiffrin memory model.
XVFI<T> XVFI: extreme video frame interpolation (Sim, Oh and Kim, ICCV 2021).

Base Classes (8)

Type Summary
DiffusionVideoSuperResolutionBase<T> Base class for DIFFUSION-based video super-resolution models.
FrameInterpolationBase<T> Base class for frame interpolation models that generate intermediate frames between existing frames.
OpticalFlowBase<T> Base class for optical flow estimation models that compute dense pixel-wise motion between frames.
VideoDenoisingBase<T> Base class for video denoising models that remove noise from video sequences.
VideoInpaintingBase<T> Base class for video inpainting models that fill in missing or masked regions in video sequences.
VideoNeuralNetworkBase<T> Base class for video-focused neural networks that can operate in both ONNX inference and native training modes.
VideoStabilizationBase<T> Base class for video stabilization models that remove camera shake from video sequences.
VideoSuperResolutionBase<T> Base class for video super-resolution models that upscale low-resolution video to higher resolution.

Interfaces (2)

Type Summary
IVideoEnhancer<T> Interface for video enhancement models that improve video quality.
IVideoModel<T> Base interface for all video AI models in AiDotNet.

Enums (4)

Type Summary
AttentionType Attention type for TimeSformer.
MiDaSVariant MiDaS model variant.
SAM2ModelSize SAM 2’s released sizes.
VideoEnhancementType Types of video enhancement operations.

Options & Configuration (104)

Type Summary
ABMEOptions Configuration options for ABME, asymmetric bilateral motion estimation for video frame interpolation (Park et al., ICCV 2021).
AMTOptions Configuration options for AMT, all-pairs multi-field transforms for frame interpolation (Li et al., CVPR 2023).
AVIDOptions Configuration options for the AVID (Audio-Visual Inpainting Diffusion) video inpainting model.
BSVDOptions Configuration options for BSVD bidirectional streaming video denoising.
BasicVSROptions Configuration options for the BasicVSR bidirectional recurrent video super-resolution model.
BasicVSRPlusPlusOptions Configuration options for the BasicVSRPlusPlus video model.
BiMVFIOptions Configuration options for BiMVFI, bidirectional motion field-guided frame interpolation (Seo, Oh and Kim, CVPR 2025).
CutieOptions Configuration options for the Cutie video segmentation model.
DAMVSROptions Configuration options for DAM-VSR, appearance and motion disentanglement for video super-resolution on Stable Video Diffusion.
DIFRINTOptions Configuration options for the DIFRINT video stabilization model.
DKMOptions Configuration options for the DKM (Dense Kernelized Matching) model.
DOVEOptions Configuration options for DOVE, efficient one-step diffusion for real-world video super-resolution.
DPFlowOptions Configuration options for the DPFlow (Dual-Path Flow) model.
DUTOptions Configuration options for the DUT (Deep Unsupervised Trajectory) video stabilization model.
DepthAnythingV2Options Configuration options for the Depth Anything V2 monocular depth model.
DualXVSROptions Configuration options for DualX-VSR, the dual axial spatial x temporal transformer for real-world video super-resolution without motion compensation.
DynamiCrafterOptions Configuration options for the DynamiCrafter video diffusion interpolation model.
E2FGVIOptions Configuration options for the E2FGVI video inpainting model.
EDVROptions Configuration options for the EDVR video model.
EMAVFIOptions Configuration options for EMAVFI, frame interpolation by extracting motion and appearance via inter-frame attention (Zhang et al., CVPR 2023).
FILMOptions Configuration options for the FILM frame interpolation model.
FLAVROptions Configuration options for FLAVR, flow-agnostic video representations for fast frame interpolation (Kalluri et al., WACV 2023).
FastDVDNetOptions Configures FastDVDnet video-denoising training and architecture settings.
FiganOptions Configuration for Figan.
FlashVSROptions Configuration options for FlashVSR, real-time diffusion-based streaming video super-resolution.
FloRNNOptions Configuration options for FloRNN (Li et al., ECCV 2022), a unidirectional video denoiser whose look-ahead recurrent module mimics a backward one.
FlowDiffuserOptions Configuration options for the FlowDiffuser model.
FlowFormerOptions Configuration options for the FlowFormer optical flow model.
FlowFormerPlusPlusOptions Configuration options for the FlowFormer++ model.
FlowLensOptions Configuration options for the FlowLens video inpainting model.
FuStaOptions Configuration options for the FuSta (Full-frame Stabilization) video stabilization model.
FuseFormerOptions Configuration options for the FuseFormer video inpainting model.
GIMMVFIOptions Configuration options for the GIMM-VFI generalizable implicit motion modeling.
GMFlowOptions Configuration options for the GMFlow optical flow model.
GaVSOptions Configuration options for the GaVS (Gaze-aware Video Stabilization) model.
IARTOptions Configuration options for IART, the implicit-resampling alignment recurrent video super-resolution transformer.
IFRNetOptions Configuration options for IFRNet, the intermediate feature refine network for efficient frame interpolation (Kong et al., CVPR 2022).
IQVFIOptions Configuration options for IQ-VFI image quality-aware video frame interpolation.
IconVSROptions Configuration options for IconVSR, BasicVSR with information refill and coupled propagation.
InternVideo2Options Configuration options for the InternVideo2 video understanding model.
InterpAnyClearerOptions Configuration options for InterpAnyClearer, distance indexing and iterative reference-based estimation for clearer anytime frame interpolation (Zhong et al., ECCV 2024).
LiteDVDNetOptions Configuration options for LiteDVDNet lightweight deep video denoising.
M2MOptions Configuration options for M2M, many-to-many splatting for video frame interpolation (Hu, Niklaus, Sclaroff and Saenko, CVPR 2022).
MGLDVSROptions Configuration options for MGLD-VSR, motion-guided latent diffusion for real-world video super-resolution.
MIAVSROptions Configuration options for MIA-VSR, the masked inter- and intra-frame attention video super-resolution transformer (Zhou et al., CVPR 2024).
McnetOptions Configuration for Mcnet.
MemFlowOptions Configuration options for the MemFlow model.
MiDaSOptions Configuration options for the MiDaS depth estimation model.
MoGOptions Configuration options for MoG motion-aware generative frame interpolation.
MoMoOptions Configuration options for MoMo momentum diffusion model for bi-directional flow.
NeuFlowV2Options Configuration options for the NeuFlow V2 model.
PSRTOptions Configuration options for PSRT-recurrent, the patch-alignment recurrent video super-resolution transformer.
PWStableNetOptions Configuration options for the PWStableNet (Pixel-Wise Stable Net) video stabilization model.
PerVFIOptions Configuration options for PerVFI perception-oriented video frame interpolation.
ProPainterOptions Configuration options for the ProPainter video inpainting model.
RAFTOptions Configuration options for the RAFT optical flow model.
RAPIDFlowOptions Configuration options for the RAPID Flow model.
RIFEOptions Configuration options for the RIFE frame interpolation model.
RPKNetOptions Configuration options for the RPKNet model.
RVMOptions Configuration options for the RVM video matting model.
RVRTOptions Configuration options for RVRT, the recurrent video restoration transformer with guided deformable attention.
RealBasicVSROptions Configuration options for RealBasicVSR, real-world video super-resolution with an image-cleaning module.
RealBasicVSRSharpOptions Configuration options for RealBasicVSR-Sharp, the adversarial training stage of RealBasicVSR.
RealESRGANOptions Configuration options for the RealESRGAN super-resolution model.
RealESRGANVideoOptions Configuration options for Real-ESRGAN applied to video, frame by frame.
RealViformerOptions Configuration options for RealViformer, the channel-attention recurrent transformer for real-world video super-resolution.
RealisVSROptions Options for RealisVSR: Wan 2.1-1.3B with the Consistency-Preserved ControlNet, trained with the High-Frequency Rectified Diffusion Loss (Zhao et al.
RoMaOptions Configuration options for the RoMa (Robust Dense Matching) model.
SAM2Options Configuration options for SAM2, SAM 2 (Ravi et al.
SEARAFTOptions Configuration options for the SEA-RAFT model.
SKFlowOptions Configuration options for the SKFlow model.
STMFNetOptions Configuration options for STMFNet spatio-temporal multi-flow network.
STTNOptions Configuration options for the STTN (Spatial-Temporal Transformer Network) video inpainting model.
SeedVROptions Options for SeedVR: the authors’ released SeedVR-7B (ByteDance-Seed/SeedVR, configs_7b/main.yaml), a window-attention MM-DiT on a causal video VAE.
ShiftNetOptions Configuration options for Shift-Net, the grouped spatial-temporal shift video restorer (Li et al., CVPR 2023), in its denoising configuration.
SlowFastOptions Configuration options for the SlowFast video model.
SoftSplatOptions Configuration options for SoftSplat, softmax splatting for video frame interpolation (Niklaus and Liu, CVPR 2020).
StabStitchOptions Configuration options for the StabStitch video stabilization model.
StableVideoSROptions Configuration options for the StableVideoSR temporal-conditioned diffusion video super-resolution model.
StreamDiffVSROptions Options for StreamDiffVSR: the authors’ released Stream-DiffVSR (jamichss/Stream-DiffVSR, Jamichsu/Stream-DiffVSR on the Hugging Face Hub).
SwinVFIOptions Configuration options for SwinVFI Swin Transformer-based video frame interpolation.
TLBVFIOptions Configuration options for TLBVFI token-level bidirectional video frame interpolation.
TTVSROptions Configuration options for TTVSR, the trajectory-aware transformer for video super-resolution.
ThreeDMFOptions Configuration options for the 3DMF (3D Motion Field) video stabilization model.
TimeSformerOptions Configuration options for the TimeSformer video model.
ToonCrafterOptions Configuration options for ToonCrafter cartoon/anime video frame interpolation.
UDVDOptions Configuration options for UDVD, unsupervised deep video denoising with a blind-spot network (Sheth et al., ICCV 2021).
UFMOptions Configuration options for the UFM (Unified Flow Matching) model.
UPRNetOptions Configuration options for UPR-Net unified pyramid recurrent network.
UniMatchOptions Configuration options for the UniMatch model.
Upscale4KAgentOptions Configuration options for Upscale4KAgent, 4KAgent (Zuo et al.
VFIMambaOptions Configuration options for VFIMamba state-space model for video frame interpolation.
VFITOptions Configuration options for VFIT, the video frame interpolation transformer (Shi et al., CVPR 2022).
VFIformerOptions Configuration options for VFIformer, video frame interpolation with a transformer (Lu et al., CVPR 2022).
VRTOptions Configuration options for the VRT video restoration model.
VideoCLIPVideoOptions Configuration options for the VideoCLIP video understanding model.
VideoEnhancementOptions<T> Configuration options for video enhancement models.
VideoFlowOptions Configuration options for the VideoFlow model.
VideoGigaGANOptions Configuration options for the VideoGigaGAN large-scale GAN for video SR.
VideoHyperparameterOptions Shared configuration for video models: action recognition, segmentation, generation, super-resolution, frame interpolation, tracking, denoising, inpainting and depth.
VideoMAEOptions Configuration options for the VideoMAE video model.
VideoModelOptions<T> Base configuration options for all video AI models.
XMemOptions Configuration options for the XMem video segmentation model.
XVFIOptions Configuration options for XVFI, extreme video frame interpolation (Sim, Oh and Kim, ICCV 2021).

Helpers & Utilities (3)

Type Summary
FiganLoss<T> FIGAN’s multi-scale perceptual objective: an L1 + VGG content loss evaluated at every scale, and combined with a very small adversarial term.
FlowWarpHelper Backward-warps an image by an optical-flow field, using tape-recorded engine operations.
McnetLoss<T> MCnet’s objective: a pixel loss plus a GRADIENT DIFFERENCE loss in image space, combined with a conditional adversarial term.