SIFT Flow: Dense Correspondence across Scenes and Its Applications
01 May 2011-IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE Computer Society)-Vol. 33, Iss: 5, pp 978-994
TL;DR: SIFT flow is proposed, a method to align an image to its nearest neighbors in a large image corpus containing a variety of scenes, where image information is transferred from the nearest neighbors to a query image according to the dense scene correspondence.
Abstract: While image alignment has been studied in different areas of computer vision for decades, aligning images depicting different scenes remains a challenging problem. Analogous to optical flow, where an image is aligned to its temporally adjacent frame, we propose SIFT flow, a method to align an image to its nearest neighbors in a large image corpus containing a variety of scenes. The SIFT flow algorithm consists of matching densely sampled, pixelwise SIFT features between two images while preserving spatial discontinuities. The SIFT features allow robust matching across different scene/object appearances, whereas the discontinuity-preserving spatial model allows matching of objects located at different parts of the scene. Experiments show that the proposed approach robustly aligns complex scene pairs containing significant spatial differences. Based on SIFT flow, we propose an alignment-based large database framework for image analysis and synthesis, where image information is transferred from the nearest neighbors to a query image according to the dense scene correspondence. This framework is demonstrated through concrete applications such as motion field prediction from a single image, motion synthesis via object transfer, satellite image registration, and face recognition.
Citations
More filters
07 Jun 2015
TL;DR: The key insight is to build “fully convolutional” networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning.
Abstract: Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, exceed the state-of-the-art in semantic segmentation. Our key insight is to build “fully convolutional” networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet [20], the VGG net [31], and GoogLeNet [32]) into fully convolutional networks and transfer their learned representations by fine-tuning [3] to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves state-of-the-art segmentation of PASCAL VOC (20% relative improvement to 62.2% mean IU on 2012), NYUDv2, and SIFT Flow, while inference takes less than one fifth of a second for a typical image.
28,225 citations
Additional excerpts
...[25] 76....
[...]
Posted Content•
TL;DR: It is shown that convolutional networks by themselves, trained end- to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation.
Abstract: Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, exceed the state-of-the-art in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a novel architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves state-of-the-art segmentation of PASCAL VOC (20% relative improvement to 62.2% mean IU on 2012), NYUDv2, and SIFT Flow, while inference takes one third of a second for a typical image.
9,803 citations
TL;DR: Fully convolutional networks (FCN) as mentioned in this paper were proposed to combine semantic information from a deep, coarse layer with appearance information from shallow, fine layer to produce accurate and detailed segmentations.
Abstract: Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build “fully convolutional” networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional networks achieve improved segmentation of PASCAL VOC (30% relative improvement to 67.2% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.
4,960 citations
21 Jul 2017
TL;DR: RefineNet is presented, a generic multi-path refinement network that explicitly exploits all the information available along the down-sampling process to enable high-resolution prediction using long-range residual connections and introduces chained residual pooling, which captures rich background context in an efficient manner.
Abstract: Recently, very deep convolutional neural networks (CNNs) have shown outstanding performance in object recognition and have also been the first choice for dense classification problems such as semantic segmentation. However, repeated subsampling operations like pooling or convolution striding in deep CNNs lead to a significant decrease in the initial image resolution. Here, we present RefineNet, a generic multi-path refinement network that explicitly exploits all the information available along the down-sampling process to enable high-resolution prediction using long-range residual connections. In this way, the deeper layers that capture high-level semantic features can be directly refined using fine-grained features from earlier convolutions. The individual components of RefineNet employ residual connections following the identity mapping mindset, which allows for effective end-to-end training. Further, we introduce chained residual pooling, which captures rich background context in an efficient manner. We carry out comprehensive experiments and set new state-of-the-art results on seven public datasets. In particular, we achieve an intersection-over-union score of 83.4 on the challenging PASCAL VOC 2012 dataset, which is the best reported result to date.
2,260 citations
Additional excerpts
...[32] RGB-D − 10....
[...]
01 Oct 2016
TL;DR: A fully convolutional architecture, encompassing residual learning, to model the ambiguous mapping between monocular images and depth maps is proposed and a novel way to efficiently learn feature map up-sampling within the network is presented.
Abstract: This paper addresses the problem of estimating the depth map of a scene given a single RGB image. We propose a fully convolutional architecture, encompassing residual learning, to model the ambiguous mapping between monocular images and depth maps. In order to improve the output resolution, we present a novel way to efficiently learn feature map up-sampling within the network. For optimization, we introduce the reverse Huber loss that is particularly suited for the task at hand and driven by the value distributions commonly present in depth maps. Our model is composed of a single architecture that is trained end-to-end and does not rely on post-processing techniques, such as CRFs or other additional refinement steps. As a result, it runs in real-time on images or videos. In the evaluation, we show that the proposed model contains fewer parameters and requires fewer training data than the current state of the art, while outperforming all approaches on depth estimation. Code and models are publicly available.
1,677 citations
Cites background or methods from "SIFT Flow: Dense Correspondence acr..."
...A second cluster of related work comprises nonparametric approaches for depth transfer [10, 13, 18, 20], which typically perform feature-based matching (e....
[...]
...Karsch et al. [11] perform warping using SIFT Flow [24], followed by a global optimization scheme, whereas Konrad et al. [12] compute a median over the retrieved depth maps followed by cross-bilateral filtering for smoothing....
[...]
...[10] perform warping using SIFT Flow [18], followed by a global optimization scheme, whereas Konrad et al....
[...]
References
More filters
20 Jun 2005
TL;DR: It is shown experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection, and the influence of each stage of the computation on performance is studied.
Abstract: We study the question of feature sets for robust visual object recognition; adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping descriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds.
31,952 citations
"SIFT Flow: Dense Correspondence acr..." refers background in this paper
...1c, the two images to match may contain object instances captured from different viewpoints, placed at different spatial locations, or imaged at different scales....
[...]
20 Sep 1999
TL;DR: Experimental results show that robust object recognition can be achieved in cluttered partially occluded images with a computation time of under 2 seconds.
Abstract: An object recognition system has been developed that uses a new class of local image features. The features are invariant to image scaling, translation, and rotation, and partially invariant to illumination changes and affine or 3D projection. These features share similar properties with neurons in inferior temporal cortex that are used for object recognition in primate vision. Features are efficiently detected through a staged filtering approach that identifies stable points in scale space. Image keys are created that allow for local geometric deformations by representing blurred image gradients in multiple orientation planes and at multiple scales. The keys are used as input to a nearest neighbor indexing method that identifies candidate object matches. Final verification of each match is achieved by finding a low residual least squares solution for the unknown model parameters. Experimental results show that robust object recognition can be achieved in cluttered partially occluded images with a computation time of under 2 seconds.
16,989 citations
"SIFT Flow: Dense Correspondence acr..." refers background in this paper
...Image alignment at the scene level is thus called scene alignment....
[...]
...In this section, we will introduce two novel applications for dense scene alignment: motion field prediction from a single image and motion synthesis via transfer of moving objects common in similar scenes....
[...]
...Ç...
[...]
...After briefly discussing how SIFT flow fits in the literature of image alignment in Section 7, we conclude the paper in Section 8....
[...]
...For example, a car moving forward can appear in a street scene with an empty road and a fish may swim in a fish tank scene....
[...]
01 Jan 1988
TL;DR: The problem the authors are addressing in Alvey Project MMI149 is that of using computer vision to understand the unconstrained 3D world, in which the viewed scenes will in general contain too wide a diversity of objects for topdown recognition techniques to work.
Abstract: The problem we are addressing in Alvey Project MMI149 is that of using computer vision to understand the unconstrained 3D world, in which the viewed scenes will in general contain too wide a diversity of objects for topdown recognition techniques to work. For example, we desire to obtain an understanding of natural scenes, containing roads, buildings, trees, bushes, etc., as typified by the two frames from a sequence illustrated in Figure 1. The solution to this problem that we are pursuing is to use a computer vision system based upon motion analysis of a monocular image sequence from a mobile camera. By extraction and tracking of image features, representations of the 3D analogues of these features can be constructed.
13,993 citations
"SIFT Flow: Dense Correspondence acr..." refers background in this paper
...For example, a car moving forward can appear in a street scene with an empty road and a fish may swim in a fish tank scene....
[...]
Proceedings Article•
24 Aug 1981TL;DR: In this paper, the spatial intensity gradient of the images is used to find a good match using a type of Newton-Raphson iteration, which can be generalized to handle rotation, scaling and shearing.
Abstract: Image registration finds a variety of applications in computer vision. Unfortunately, traditional image registration techniques tend to be costly. We present a new image registration technique that makes use of the spatial intensity gradient of the images to find a good match using a type of Newton-Raphson iteration. Our technique is taster because it examines far fewer potential matches between the images than existing techniques Furthermore, this registration technique can be generalized to handle rotation, scaling and shearing. We show how our technique can be adapted tor use in a stereo vision system.
12,944 citations
TL;DR: In this paper, a method for finding the optical flow pattern is presented which assumes that the apparent velocity of the brightness pattern varies smoothly almost everywhere in the image, and an iterative implementation is shown which successfully computes the Optical Flow for a number of synthetic image sequences.
10,727 citations