- Stereo videos scraped from Youtube
- Driving stereo sequences like Kitti
- Another driving sequence
- Synthetic stereo videos with ground-truth disparity maps
- Large scale high-res stereo data from zed camera
- Urban stereo dataset
- Visual inertial stereo dataset with fish-eye lens
- Synthetic Indoor stereo video dataset
- Plenoptic and stereo camera video dataset for odometry evaluation
- Cityscapes stereo data
- Kitti stereo data
- Underwater stereo images
- Sintel dataset
- Middlebury stereo data
- Underwater stereo images
- High speed stereo video dataset (480 fps)
- https://arpg.github.io/oivio//
- stereo images data
- Open loris scene dataset
- 3D movie dataset from INRIA
- ETH3D dataset
- IIITH Stereoscopic 3D data: Stereo data collected for stereo video quality assessment
- Another stereoscopic 3D data: Stereo data collected for stereo video quality assessment
There are mainly 4 datasets:
- Raytrix 5x5 data: This one contains 5x5 3 video sequences with 5x5 angular resolution data. Has small baseline. But the spatial resolution is quite high
- Hybrid LF video dataset: A video data captured using Lytro. Small spatial resolution. angular resolution is 8x8
- X-fields dataset: Has about 8 videos. 3 videos have 3x3x3 views, i.e. 3 frames with each 3x3 angualr views. Another set of videos has 5x5x5 views, i.e. 5 frames each with 5x5 angular views.
- Camera grid dataset
- Mobile phone stereo examples from DU2Net
- ICCV 2017 : Coherent Online Video Style Transfer
- Siggraph Asia 2016 : Temporally Coherent Completion of Dynamic Video
- NeurIPS 2018 : Video-to-Video Synthesis; Webpage
- IEEE RAL 2020: Don't Forget The Past: Recurrent Depth Estimation from Monocular Video
- Unpublished : Robust Consistent Video Depth Estimation
- CVPR 2019 - Deep Video Inpainting: Interesting paper
- Unpublished - World-Consistent Video-to-Video Synthesis; Webpage
- ICCV 2019 - Onion-Peel Networks for Deep Video Completion: uses an attention mechanism similar to those in NLP like BERT etc. There is not really anything about recurrence. This is more like Video denoising or Video SR papers.
- ICCV 2019 - Exploiting temporal consistency for real-time video depth estimation
- CVPR 2019 - Semantic Image Synthesis with Spatially-Adaptive Normalization: Not a video paper, but very interesting paper. This generates an image given a semantic map. It has a generator which takes input a noise vector from Gaussian distribution. Then it modulates the features map like batch-norm but the mean and variance vectors are derived from the semantic map.
- CVPR 2019 - Single-frame Regularization for Temporally Stable CNNs: Somewhat an interesting paper which tries to ensure local temporal consistency for networks trained on single images.
- Learning Single Camera Depth Estimation using Dual-Pixels
- DU2-net: Learning Depth Estimation from Dual-Cameras and Dual-Pixels
- DeepLens: Shallow Depth of Field from a Single Image has Code and dataset is proposed but not released
- Deep Sparse Light Field Refocusing: Pretty interesting paper as in it does direct refocusing from sparse light fields
- Siggraph Asia 2016 - Learning-Based View Synthesis for Light Field Cameras; Webpage; Dataset; Code
- CVPR 2017 - Light Field Reconstruction Using Deep Convolutional Network on EPI: Input 3x3 sparse views. The network takes the EPI images as input and does the upsampling. A supervised framework.
- ICCV 2017 - Learning to Synthesize a 4D RGBD Light Field from a Single Image: Train the network on lots of flower images. Then reconstruct light field from a single image. Disparity based rendering is used. Non-lambertian effects are synthesized using a residual block using the supervision.
- Siggraph 2018 - Stereo Magnification: Learning view synthesis using multiplane images
- CVPRW 2017 - Compressive Light Field Reconstructions using Deep Learning
- ECCV 2018 - Learning to capture light fields through a coded aperture camera
- ICCV 2017 - Neural EPI-volume networks for shape from light field
- ECCV 2018 - End-to-end view synthesis for light field imaging with pseudo 4DCNN
- CVPR 2018 - Enhancing the spatial resolution of stereo images using a parallax prior
- ICCV 2019 - Extreme View Synthesis: It's like the stereo magnification paper
- Selected Topics in Circuits and Systems - Light Field Image Compression Using Generative Adversarial Network-Based View Synthesis
- TIP - Light Field Super-Resolution using a Low-Rank Prior and Deep Convolutional Neural Networks
- ACCV 2018 - Dense light field reconstruction from sparse sampling using residual network
- Siggraph Asia 20 - Synthesizing light field from a single image with variable MPI and two network fusion
- CVPR 2015 - Light Field from Micro-baseline Image Pair: A traditional, non-learning based method which first predicts the stereo disparity map and then renders the light field
- Siggraph 2013 - Joint View Expansion and Filtering for Automultiscopic 3D Displays: No code is available; however test images are available Webpage
- TIP 2018 Spatial and angular resolution enhancement of light fields using convolutional neural networks
- [Light Field Super-resolution via Attention-Guided Fusion of Hybrid Lenses]
- Siggraph 2017 - Light field video capture using a learning-based hybrid imaging system
- 3DTV at home: eulerian-lagrangian stereo-to-multiview conversion
- CGF journal 2020 - Single Sensor Compressive Light Field Video Camera
- unpublished - 5D Light Field Synthesis from a Monocular Video
- Sig Asia 2020 - X-Fields: Implicit Neural View-, Light- and Time-Image Interpolation : Contains LF video dataset also has Code and Dataset
- ICCV 2021 - SeLFVi: Self-supervised Light Field Video Reconstruction from Stereo Video: Code, Webpage and Supplementary Material
- Perceptual evaluation of light field image
- Light Field Image Quality Assessment: An Overview
- Defocus evaluation