Today, visual recognition systems are still rarely employed in robotics applications. Perhaps one of the main reasons for this is the lack of demanding benchmarks that mimic such scenarios. In this paper, we take advantage of our autonomous driving platform to develop novel challenging benchmarks for the tasks of stereo, optical flow, visual odometry/SLAM and 3D object detection. Our recording platform is equipped with four high resolution video cameras, a Velodyne laser scanner and a state-of-the-art localization system. Our benchmarks comprise 389 stereo and optical flow image pairs, stereo visual odometry sequences of 39.2 km length, and more than 200k 3D object annotations captured in cluttered scenarios (up to 15 cars and 30 pedestrians are visible per image). Results from state-of-the-art algorithms reveal that methods ranking high on established datasets such as Middlebury perform below average when being moved outside the laboratory to the real world. Our goal is to reduce this bias by providing challenging benchmarks with novel difficulties to the computer vision community. Our benchmarks are available online at: www.cvlibs.net/datasets/kitti

/pdf/are-we-ready-for-autonomous-driving-the-kitti-vision-zygefho352.pdf

Are we ready for autonomous driving? The KITTI vision benchmark suite

We investigate architectures of discriminatively trained deep Convolutional Networks (ConvNets) for action recognition in video. The challenge is to capture the complementary information on appearance from still frames and motion between frames. We also aim to generalise the best performing hand-crafted features within a data-driven learning framework.

Our contribution is three-fold. First, we propose a two-stream ConvNet architecture which incorporates spatial and temporal networks. Second, we demonstrate that a ConvNet trained on multi-frame dense optical flow is able to achieve very good performance in spite of limited training data. Finally, we show that multitask learning, applied to two different action classification datasets, can be used to increase the amount of training data and improve the performance on both. Our architecture is trained and evaluated on the standard video actions benchmarks of UCF-101 and HMDB-51, where it is competitive with the state of the art. It also exceeds by a large margin previous attempts to use deep nets for video classification.

https://proceedings.neurips.cc/paper/2014/file/00ec53c4682d36f5c4359f4ae7bd7ba1-Paper.pdf

Two-Stream Convolutional Networks for Action Recognition in Videos

The goal of precipitation nowcasting is to predict the future rainfall intensity in a local region over a relatively short period of time. Very few previous studies have examined this crucial and challenging weather forecasting problem from the machine learning perspective. In this paper, we formulate precipitation nowcasting as a spatiotemporal sequence forecasting problem in which both the input and the prediction target are spatiotemporal sequences. By extending the fully connected LSTM (FC-LSTM) to have convolutional structures in both the input-to-state and state-to-state transitions, we propose the convolutional LSTM (ConvLSTM) and use it to build an end-to-end trainable model for the precipitation nowcasting problem. Experiments show that our ConvLSTM network captures spatiotemporal correlations better and consistently outperforms FC-LSTM and the state-of-the-art operational ROVER algorithm for precipitation nowcasting.

Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting

Models based on deep convolutional networks have dominated recent image interpretation tasks; we investigate whether models which are also recurrent, or “temporally deep”, are effective for tasks involving sequences, visual and otherwise. We develop a novel recurrent convolutional architecture suitable for large-scale visual learning which is end-to-end trainable, and demonstrate the value of these models on benchmark video recognition tasks, image description and retrieval problems, and video narration challenges. In contrast to current models which assume a fixed spatio-temporal receptive field or simple temporal averaging for sequential processing, recurrent convolutional models are “doubly deep” in that they can be compositional in spatial and temporal “layers”. Such models may have advantages when target concepts are complex and/or training data are limited. Learning long-term dependencies is possible when nonlinearities are incorporated into the network state updates. Long-term RNN models are appealing in that they directly can map variable-length inputs (e.g., video frames) to variable length outputs (e.g., natural language text) and can model complex temporal dynamics; yet they can be optimized with backpropagation. Our recurrent long-term models are directly connected to modern visual convnet models and can be jointly trained to simultaneously learn temporal dynamics and convolutional perceptual representations. Our results show such models have distinct advantages over state-of-the-art models for recognition or generation which are separately defined and/or optimized.

/pdf/long-term-recurrent-convolutional-networks-for-visual-3znyib4gwu.pdf

Long-term recurrent convolutional networks for visual recognition and description

Humans perceive the three-dimensional structure of the world with apparent ease. However, despite all of the recent advances in computer vision research, the dream of having a computer interpret an image at the same level as a two-year old remains elusive. Why is computer vision such a challenging problem and what is the current state of the art? Computer Vision: Algorithms and Applications explores the variety of techniques commonly used to analyze and interpret images. It also describes challenging real-world applications where vision is being successfully used, both for specialized applications such as medical imaging, and for fun, consumer-level tasks such as image editing and stitching, which students can apply to their own personal photos and videos. More than just a source of recipes, this exceptionally authoritative and comprehensive textbook/reference also takes a scientific approach to basic vision problems, formulating physical models of the imaging process before inverting them to produce descriptions of a scene. These problems are also analyzed using statistical models and solved using rigorous engineering techniques Topics and features: structured to support active curricula and project-oriented courses, with tips in the Introduction for using the book in a variety of customized courses; presents exercises at the end of each chapter with a heavy emphasis on testing algorithms and containing numerous suggestions for small mid-term projects; provides additional material and more detailed mathematical topics in the Appendices, which cover linear algebra, numerical techniques, and Bayesian estimation theory; suggests additional reading at the end of each chapter, including the latest research in each sub-field, in addition to a full Bibliography at the end of the book; supplies supplementary course material for students at the associated website, http://szeliski.org/Book/. Suitable for an upper-level undergraduate or graduate-level course in computer science or engineering, this textbook focuses on basic techniques that work under real-world conditions and encourages students to push their creative boundaries. Its design and exposition also make it eminently suitable as a unique reference to the fundamental techniques and current research literature in computer vision.

/pdf/computer-vision-algorithms-and-applications-25dn6wu83j.pdf

Computer Vision: Algorithms and Applications

We suggest a variational method for the joint estimation of optic flow and the segmentation of the image into regions of similar motion. It makes use of the level set framework following the idea of motion competition, which is extended to non-parametric motion. Moreover, we automatically determine an appropriate initialization and the number of regions by means of recursive two-phase splits with higher order region models. The method is further extended to the spatiotemporal setting and the use of additional cues like the gray value or color for the segmentation. It need not fear a quantitative comparison to pure optic flow estimation techniques: For the popular Yosemite sequence with clouds we obtain the currently most accurate result. We further uncover a mistake in the ground truth. Coarsely correcting this, we get an average angular error below 1 degree.

Variational Motion Segmentation with Level Sets

/pdf/variational-motion-segmentation-with-level-sets-4r4op2c4z6.pdf

Variational motion segmentation with level sets

There are two popular ways to implement anisotropic diffusion filters with a diffusion tensor: Explicit finite difference schemes are simple but become inefficient due to severe time step size restrictions, while semi-implicit schemes are more efficient but require to solve large linear systems of equations. In our paper we present a novel class of algorithms that combine the advantages of both worlds: They are based on simple explicit schemes, while beingmore efficient than semi-implicit approaches. These so-called fast explicit diffusion (FED) schemes perform cycles of explicit schemeswith varying time step sizes that may violate the stability restriction in up to 50 percent of all cases. FED schemes can be motivated froma decomposition of box filters in terms of explicit schemes for linear diffusion problems. Experiments demonstrate the advantages of the FEDapproach for timedependent (parabolic) image enhancement problems as well as for steady state (elliptic) image compression tasks. In the latter case FED schemes are speeded up substantially by embedding them in a cascadic coarse-to-fine approach.

/pdf/from-box-filtering-to-fast-explicit-diffusion-4janeavuut.pdf

From box filtering to fast explicit diffusion

This paper deals with establishing relations between a number of widely-used nonlinear filters for digital image processing. We cover robust statistical estimation with (local) M-estimators, local mode filtering in image or histogram space, bilateral filtering, nonlinear diusion, and regularisation approaches. Although these methods originate in dierent mathematical theories, we show that their implementation reveals a highly similar structure. We demonstrate that all these methods can be cast into a unified framework of functional minimisation combining nonlocal data and nonlocal smoothness terms. This unification contributes to a better understanding of the individual methods, and it opens the way to new techniques combining the advantages of known filters.

/pdf/on-robust-estimation-and-smoothing-with-spatial-and-tonal-2eokg8w4ng.pdf

On robust estimation and smoothing with spatial and tonal kernels

Optic fow describes the displacement field in an image sequence. Its reliable computation constitutes one of the main challenges in computer vision, and variational methods belong to the most successful techniques for achieving this goal. Variational methods recover the optic flow field as a minimiser of a suitable energy functional that involves data and smoothness terms. In this paper we present a survey on different model assumptions for each of these terms and illustrate their impact by experiments. We restrict ourselves to rotationally invariant convex functionals with a linearised data term. Such models are appropriate for small displacements. Regarding the data term, constancy assumptions on the brightness, the gradient, the Hessian, the gradient magnitude, the Laplacian, and the Hessian determinant are investigated. Local integration and nonquadratic penalisation are considered in order to improve robustness under noise. With respect to the smoothness term, we review a recent taxonomy that links regularisers to diffusion processes. It allows to distinguish five types of regularisation strategies: homogeneous, isotropic image-driven, anisotropic image-driven, isotropic flow-driven, and anisotropic flow-driven. All these regularisations can be performed either in the spatial or the spatiotemporal domain. After discussing well-posedness results for convex optic flow functionals, we sketch some numerical ideas in order to achieve realtime performance on a standard PC by means of multigrid methods, and we survey a simple and intuitive confidence measure.

/pdf/a-survey-on-variational-optic-flow-methods-for-small-2fiicpty91.pdf

Andrés Bruhn

Papers

Variational Motion Segmentation with Level Sets

Variational motion segmentation with level sets

From box filtering to fast explicit diffusion

On robust estimation and smoothing with spatial and tonal kernels

A survey on variational optic flow methods for small displacements