DeepFlow: Large Displacement Optical Flow with Deep Matching

By Philippe Weinzaepfel, Jérôme Revaud, Zaïd Harchaoui, Cordelia Schmid. 2013

1,118 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

What makes ImageNet good for transfer learning? (2016, methodology): “Using ImageNet pre-trained CNN features, impressive results have been obtained on several image classification datasets [13, 35], object detection [16, 39], action recognition [41], human pose estimation [9], image segmentation [10], optical flow [47], image captioning [12, 22] and others [26].”
Two-Stream Convolutional Networks for Action Recognition in Videos (2014, background): “The latter is probably too challenging, and might require architectural changes (for example, a combination with the deep matching approach of [30]).”
Deep Clustering for Unsupervised Learning of Visual Features (2018, background): “Pre-trained convolutional neural networks, or convnets, have become the building blocks in most computer vision applications [8, 9, 50, 65].”
FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks (2016, methodology): “91% † DeepFlow [32] 80.18 % 29 / 65 81.89 % EpicFlow [22] 78 .”
FlowNet: Learning Optical Flow with Convolutional Networks (2015, methodology): “The successive work of [30] termed EpicFlow has put even more emphasis on the quality of sparse matching as the matches from [35] are merely interpolated to dense flow fields while respecting image boundaries.”
DeepDriving: Learning Affordance for Direct Perception in Autonomous Driving (2015, methodology): “For depth map estimation, DeepFlow [20] uses ConvNets to achieve very good results for driving scene images on the KITTI dataset [6].”
Unsupervised Learning of Edges (2015, methodology): “The only input to our approach is semi-dense matching results from [42].”
Spatio-temporal video autoencoder with differentiable memory (2015, methodology): “Our approach is partially related to optical ﬂow estimation works like DeepFlow Weinzaepfel et al. (2013) and FlowNet Fischer et al. (2015).”
Cross Pixel Optical Flow Similarity for Self-Supervised Learning (2018, positive): “Epic-Flow [51] + Deep-Matching [52] was used to compute opticalflow for these frames.”
Dense Optical Flow Prediction from a Static Image (2015, methodology): “We chose the publicly available implementation of DeepFlow [27] to compute optical flow.”
Recurrent Pixel Embedding for Instance Grouping (2017, background): “The successes of deep convolutional neural nets (CNNs) at image classification has spawned a flurry of work in computer vision on adapting these models to pixel-level image understanding tasks, such as boundary detection [1, 89, 63], semantic segmentation [59, 10, 45], optical flow [86, 20], and pos”
segDeepM: Exploiting segmentation and context in deep neural networks for object detection (2015, neutral): “They have been applied to a variety of general vision problems, such as recognition [15, 9], segmentation [11], stereo [18], flow [24], and even text-from-image generation [13], consistently outperforming past work.”
No More Discrimination: Cross City Adaptation of Road Scene Segmenters (2017, methodology): “4, given an image pair of the same location but across different times, we first apply DeepMatching [36] to relate pixels within each image pair.”
Video Frame Interpolation via Adaptive Separable Convolution (2017, methodology): “Our research is inspired by the success of applying deep learning to optical ﬂow estimation [2, 12, 16, 19, 50, 51, 52], artistic style transfer [17, 22, 28], and image enhancement [6, 9, 10, 46, 47, 55, 57, 62, 65].”
PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume (2017, methodology): “To address this issue, Brox and Malik [10] embed feature matching into the variational framework, which is further improved by follow-up methods [50, 56].”
Dynamic FAUST: Registering Human Bodies in Motion (2017, methodology): “We compute dense matches [52] in 2D between each map At and a reference map A (red lines).”
Learning to Extract Motion from Videos in Convolutional Neural Networks (2016, background): “(3) We show how to enforce strict rotation invariance within a CNN through weight-sharing, and demonstrate signiﬁcant beneﬁt for training on a small dataset with no need for data augmentation.”