Showing papers by "French Institute for Research in Computer Science and Automation published in 2013"

Action Recognition with Improved Trajectories

Abstract: Recently dense trajectories were shown to be an efficient video representation for action recognition and achieved state-of-the-art results on a variety of datasets. This paper improves their performance by taking into account camera motion to correct them. To estimate camera motion, we match feature points between frames using SURF descriptors and dense optical flow, which are shown to be complementary. These matches are, then, used to robustly estimate a homography with RANSAC. Human motion is in general different from camera motion and generates inconsistent matches. To improve the estimation, a human detector is employed to remove these matches. Given the estimated camera motion, we remove trajectories consistent with it. We also use this estimation to cancel out camera motion from the optical flow. This significantly improves motion-based descriptors, such as HOF and MBH. Experimental results on four challenging action datasets (i.e., Hollywood2, HMDB51, Olympic Sports and UCF50) significantly outperform the current state of the art.

Dense Trajectories and Motion Boundary Descriptors for Action Recognition

Abstract: This paper introduces a video representation based on dense trajectories and motion boundary descriptors. Trajectories capture the local motion information of the video. A dense representation guarantees a good coverage of foreground motion as well as of the surrounding context. A state-of-the-art optical flow algorithm enables a robust and efficient extraction of dense trajectories. As descriptors we extract features aligned with the trajectories to characterize shape (point coordinates), appearance (histograms of oriented gradients) and motion (histograms of optical flow). Additionally, we introduce a descriptor based on motion boundary histograms (MBH) which rely on differential optical flow. The MBH descriptor shows to consistently outperform other state-of-the-art descriptors, in particular on real-world videos that contain a significant amount of camera motion. We evaluate our video representation in the context of action classification on nine datasets, namely KTH, YouTube, Hollywood2, UCF sports, IXMAS, UIUC, Olympic Sports, UCF50 and HMDB51. On all datasets our approach outperforms current state-of-the-art results.

Image Classification with the Fisher Vector: Theory and Practice

Abstract: A standard approach to describe an image for classification and retrieval purposes is to extract a set of local patch descriptors, encode them into a high dimensional vector and pool them into an image-level signature The most common patch encoding strategy consists in quantizing the local descriptors into a finite set of prototypical elements This leads to the popular Bag-of-Visual words representation In this work, we propose to use the Fisher Kernel framework as an alternative patch encoding strategy: we describe patches by their deviation from an "universal" generative Gaussian mixture model This representation, which we call Fisher vector has many advantages: it is efficient to compute, it leads to excellent results even with efficient linear classifiers, and it can be compressed with a minimal loss of accuracy using product quantization We report experimental results on five standard datasets--PASCAL VOC 2007, Caltech 256, SUN 397, ILSVRC 2010 and ImageNet10K--with up to 9M images and 10K classes, showing that the FV framework is a state-of-the-art patch encoding technique

Advances in Spectral-Spatial Classification of Hyperspectral Images

Abstract: Recent advances in spectral-spatial classification of hyperspectral images are presented in this paper. Several techniques are investigated for combining both spatial and spectral information. Spatial information is extracted at the object (set of pixels) level rather than at the conventional pixel level. Mathematical morphology is first used to derive the morphological profile of the image, which includes characteristics about the size, orientation, and contrast of the spatial structures present in the image. Then, the morphological neighborhood is defined and used to derive additional features for classification. Classification is performed with support vector machines (SVMs) using the available spectral information and the extracted spatial information. Spatial postprocessing is next investigated to build more homogeneous and spatially consistent thematic maps. To that end, three presegmentation techniques are applied to define regions that are used to regularize the preliminary pixel-wise thematic map. Finally, a multiple-classifier (MC) system is defined to produce relevant markers that are exploited to segment the hyperspectral image with the minimum spanning forest algorithm. Experimental results conducted on three real hyperspectral images with different spatial and spectral resolutions and corresponding to various contexts are presented. They highlight the importance of spectral-spatial strategies for the accurate classification of hyperspectral images and validate the proposed methods.

Ontology Matching: State of the Art and Future Challenges

Abstract: After years of research on ontology matching, it is reasonable to consider several questions: is the field of ontology matching still making progress? Is this progress significant enough to pursue further research? If so, what are the particularly promising directions? To answer these questions, we review the state of the art of ontology matching and analyze the results of recent ontology matching evaluations. These results show a measurable improvement in the field, the speed of which is albeit slowing down. We conjecture that significant improvements can be obtained only by addressing important challenges for ontology matching. We present such challenges with insights on how to approach them, thereby aiming to direct research into the most promising tracks and to facilitate the progress of the field.

API design for machine learning software: experiences from the scikit-learn project

Abstract: Scikit-learn is an increasingly popular machine learning li- brary. Written in Python, it is designed to be simple and efficient, accessible to non-experts, and reusable in various contexts. In this paper, we present and discuss our design choices for the application programming interface (API) of the project. In particular, we describe the simple and elegant interface shared by all learning and processing units in the library and then discuss its advantages in terms of composition and reusability. The paper also comments on implementation details specific to the Python ecosystem and analyzes obstacles faced by users and developers of the library.

DeepFlow: Large Displacement Optical Flow with Deep Matching

Abstract: Optical flow computation is a key component in many computer vision systems designed for tasks such as action detection or activity recognition. However, despite several major advances over the last decade, handling large displacement in optical flow remains an open problem. Inspired by the large displacement optical flow of Brox and Malik, our approach, termed Deep Flow, blends a matching algorithm with a variational approach for optical flow. We propose a descriptor matching algorithm, tailored to the optical flow problem, that allows to boost performance on fast motions. The matching algorithm builds upon a multi-stage architecture with 6 layers, interleaving convolutions and max-pooling, a construction akin to deep convolutional nets. Using dense sampling, it allows to efficiently retrieve quasi-dense correspondences, and enjoys a built-in smoothing effect on descriptors matches, a valuable asset for integration into an energy minimization framework for optical flow estimation. Deep Flow efficiently handles large displacements occurring in realistic videos, and shows competitive performance on optical flow benchmarks. Furthermore, it sets a new state-of-the-art on the MPI-Sintel dataset.

Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME.

Abstract: This review encompasses the most important advances in liver functions and hepatotoxicity and analyzes which mechanisms can be studied in vitro. In a complex architecture of nested, zonated lobules, the liver consists of approximately 80 % hepatocytes and 20 % non-parenchymal cells, the latter being involved in a secondary phase that may dramatically aggravate the initial damage. Hepatotoxicity, as well as hepatic metabolism, is controlled by a set of nuclear receptors (including PXR, CAR, HNF-4α, FXR, LXR, SHP, VDR and PPAR) and signaling pathways. When isolating liver cells, some pathways are activated, e.g., the RAS/MEK/ERK pathway, whereas others are silenced (e.g. HNF-4α), resulting in up- and downregulation of hundreds of genes. An understanding of these changes is crucial for a correct interpretation of in vitro data. The possibilities and limitations of the most useful liver in vitro systems are summarized, including three-dimensional culture techniques, co-cultures with non-parenchymal cells, hepatospheres, precision cut liver slices and the isolated perfused liver. Also discussed is how closely hepatoma, stem cell and iPS cell-derived hepatocyte-like-cells resemble real hepatocytes. Finally, a summary is given of the state of the art of liver in vitro and mathematical modeling systems that are currently used in the pharmaceutical industry with an emphasis on drug metabolism, prediction of clearance, drug interaction, transporter studies and hepatotoxicity. One key message is that despite our enthusiasm for in vitro systems, we must never lose sight of the in vivo situation. Although hepatocytes have been isolated for decades, the hunt for relevant alternative systems has only just begun.

Towards Understanding Action Recognition

Abstract: Although action recognition in videos is widely studied, current methods often fail on real-world datasets. Many recent approaches improve accuracy and robustness to cope with challenging video sequences, but it is often unclear what affects the results most. This paper attempts to provide insights based on a systematic performance evaluation using thoroughly-annotated data of human actions. We annotate human Joints for the HMDB dataset (J-HMDB). This annotation can be used to derive ground truth optical flow and segmentation. We evaluate current methods using this dataset and systematically replace the output of various algorithms with ground truth. This enables us to discover what is important - for example, should we work on improving flow algorithms, estimating human bounding boxes, or enabling pose estimation? In summary, we find that high-level pose features greatly outperform low/mid level features, in particular, pose over time is critical, but current pose estimation algorithms are not yet reliable enough to provide this information. We also find that the accuracy of a top-performing action recognition framework can be greatly increased by refining the underlying low/mid level features, this suggests it is important to improve optical flow and human detection algorithms. Our analysis and J-HMDB dataset should facilitate a deeper understanding of action recognition algorithms.

High-order CFD methods: Current status and perspective

Abstract: After several years of planning, the 1st International Workshop on High-Order CFD Methods was successfully held in Nashville, Tennessee, on January 7-8, 2012, just before the 50th Aerospace Sciences Meeting. The American Institute of Aeronautics and Astronautics, the Air Force Office of Scientific Research, and the German Aerospace Center provided much needed support, financial and moral. Over 70 participants from all over the world across the research spectrum of academia, government labs, and private industry attended the workshop. Many exciting results were presented. In this review article, the main motivation and major findings from the workshop are described. Pacing items requiring further effort are presented. © 2013 John Wiley & Sons, Ltd.

Experimental demonstration of long-distance continuous-variable quantum key distribution

Abstract: Researchers demonstrate continuous-variable quantum key distribution over 80 km of optical fibre. They develop a detection system based on homodyne detectors to achieve an unprecedented level of stability and implement new codes tailored to perform secure communication optimally in a range of signal-to-noise ratios corresponding to long distances.

Label-Embedding for Attribute-Based Classification

Abstract: Attributes are an intermediate representation, which enables parameter sharing between classes, a must when training data is scarce. We propose to view attribute-based image classification as a label-embedding problem: each class is embedded in the space of attribute vectors. We introduce a function which measures the compatibility between an image and a label embedding. The parameters of this function are learned on a training set of labeled samples to ensure that, given an image, the correct classes rank higher than the incorrect ones. Results on the Animals With Attributes and Caltech-UCSD-Birds datasets show that the proposed framework outperforms the standard Direct Attribute Prediction baseline in a zero-shot learning scenario. The label embedding framework offers other advantages such as the ability to leverage alternative sources of information in addition to attributes (e.g. class hierarchies) or to transition smoothly from zero-shot learning to learning with large quantities of data.

Assemblathon 2: evaluating de novo methods of genome assembly in three vertebrate species

Abstract: Background: The process of generating raw genome sequence data continues to become cheaper, faster, and more accurate. However, assembly of such data into high-quality, finished genome sequences remains challenging. Many genome assembly tools are available, but they differ greatly in terms of their performance (speed, scalability, hardware requirements, acceptance of newer read technologies) and in their final output (composition of assembled sequence). More importantly, it remains largely unclear how to best assess the quality of assembled genome sequences. The Assemblathon competitions are intended to assess current state-of-the-art methods in genome assembly. Results: In Assemblathon 2, we provided a variety of sequence data to be assembled for three vertebrate species (a bird, a fish, and snake). This resulted in a total of 43 submitted assemblies from 21 participating teams. We evaluated these assemblies using a combination of optical map data, Fosmid sequences, and several statistical methods. From over 100 different metrics, we chose ten key measures by which to assess the overall quality of the assemblies. (Continued on next page)

Deterministically encoding quantum information using 100-photon Schrödinger cat states.

Abstract: In contrast to a single quantum bit, an oscillator can store multiple excitations and coherences provided one has the ability to generate and manipulate complex multiphoton states. We demonstrate multiphoton control by using a superconducting transmon qubit coupled to a waveguide cavity resonator with a highly ideal off-resonant coupling. This dispersive interaction is much greater than decoherence rates and higher-order nonlinearities to allow simultaneous manipulation of hundreds of photons. With a tool set of conditional qubit-photon logic, we mapped an arbitrary qubit state to a superposition of coherent states, known as a "cat state." We created cat states as large as 111 photons and extended this protocol to create superpositions of up to four coherent states. This control creates a powerful interface between discrete and continuous variable quantum computation and could enable applications in metrology and quantum information processing.

Why3: where programs meet provers

Abstract: We present Why3, a tool for deductive program verification, and WhyML, its programming and specification language. WhyML is a first-order language with polymorphic types, pattern matching, and inductive predicates. Programs can make use of record types with mutable fields, type invariants, and ghost code. Verification conditions are discharged by Why3 with the help of various existing automated and interactive theorem provers. To keep verification conditions tractable and comprehensible, WhyML imposes a static control of aliases that obviates the use of a memory model. A user can write WhyML programs directly and get correct-by-construction OCaml programs via an automated extraction mechanism. WhyML is also used as an intermediate language for the verification of C, Java, or Ada programs. We demonstrate the benefits of Why3 and WhyML on non-trivial examples of program verification.

Action and Event Recognition with Fisher Vectors on a Compact Feature Set

Abstract: Action recognition in uncontrolled video is an important and challenging computer vision problem. Recent progress in this area is due to new local features and models that capture spatio-temporal structure between local features, or human-object interactions. Instead of working towards more complex models, we focus on the low-level features and their encoding. We evaluate the use of Fisher vectors as an alternative to bag-of-word histograms to aggregate a small set of state-of-the-art low-level descriptors, in combination with linear classifiers. We present a large and varied set of evaluations, considering (i) classification of short actions in five datasets, (ii) localization of such actions in feature-length movies, and (iii) large-scale recognition of complex events. We find that for basic action recognition and localization MBH features alone are enough for state-of-the-art performance. For complex events we find that SIFT and MFCC features provide complementary cues. On all three problems we obtain state-of-the-art results, while using fewer features and less complex models.

Chaste: an open source C++ library for computational physiology and biology.

Abstract: Chaste - Cancer, Heart And Soft Tissue Environment - is an open source C++ library for the computational simulation of mathematical models developed for physiology and biology. Code development has been driven by two initial applications: cardiac electrophysiology and cancer development. A large number of cardiac electrophysiology studies have been enabled and performed, including high-performance computational investigations of defibrillation on realistic human cardiac geometries. New models for the initiation and growth of tumours have been developed. In particular, cell-based simulations have provided novel insight into the role of stem cells in the colorectal crypt. Chaste is constantly evolving and is now being applied to a far wider range of problems. The code provides modules for handling common scientific computing components, such as meshes and solvers for ordinary and partial differential equations (ODEs/PDEs). Re-use of these components avoids the need for researchers to 're-invent the wheel' with each new project, accelerating the rate of progress in new applications. Chaste is developed using industrially-derived techniques, in particular test-driven development, to ensure code quality, re-use and reliability. In this article we provide examples that illustrate the types of problems Chaste can be used to solve, which can be run on a desktop computer. We highlight some scientific studies that have used or are using Chaste, and the insights they have provided. The source code, both for specific releases and the development version, is available to download under an open source Berkeley Software Distribution (BSD) licence at http://www.cs.ox.ac.uk/chaste, together with details of a mailing list and links to documentation and tutorials.

Better Exploiting Motion for Better Action Recognition

Abstract: Several recent works on action recognition have attested the importance of explicitly integrating motion characteristics in the video description. This paper establishes that adequately decomposing visual motion into dominant and residual motions, both in the extraction of the space-time trajectories and for the computation of descriptors, significantly improves action recognition algorithms. Then, we design a new motion descriptor, the DCS descriptor, based on differential motion scalar quantities, divergence, curl and shear features. It captures additional information on the local motion patterns enhancing results. Finally, applying the recent VLAD coding technique proposed in image retrieval provides a substantial improvement for action recognition. Our three contributions are complementary and lead to outperform all reported results by a significant margin on three challenging datasets, namely Hollywood 2, HMDB51 and Olympic Sports.

The Tapenade automatic differentiation tool: Principles, model, and specification

Abstract: Tapenade is an Automatic Differentiation (AD) tool which, given a Fortran or C code that computes a function, creates a new code that computes its tangent or adjoint derivatives. Tapenade puts particular emphasis on adjoint differentiation, which computes gradients at a remarkably low cost. This article describes the principles of Tapenade, a subset of the general principles of AD. We motivate and illustrate with examples the AD model of Tapenade, that is, the structure of differentiated codes and the strategies used to make them more efficient. Along with this informal description, we formally specify this model by means of data-flow equations and rules of Operational Semantics, making this the reference specification of the tangent and adjoint modes of Tapenade. One benefit we expect from this formal specification is the capacity to formally study the AD model itself, especially for the adjoint mode and its sophisticated strategies. This article also describes the architectural choices of the implementation of Tapenade. We describe the current performance of Tapenade on a set of codes that include industrial-size applications. We present the extensions of the tool that are planned in a foreseeable future, deriving from our ongoing research on AD.

Distance-Based Image Classification: Generalizing to New Classes at Near-Zero Cost

Abstract: We study large-scale image classification methods that can incorporate new classes and training images continuously over time at negligible cost. To this end, we consider two distance-based classifiers, the k-nearest neighbor (k-NN) and nearest class mean (NCM) classifiers, and introduce a new metric learning approach for the latter. We also introduce an extension of the NCM classifier to allow for richer class representations. Experiments on the ImageNet 2010 challenge dataset, which contains over 106 training images of 1,000 classes, show that, surprisingly, the NCM classifier compares favorably to the more flexible k-NN classifier. Moreover, the NCM performance is comparable to that of linear SVMs which obtain current state-of-the-art performance. Experimentally, we study the generalization performance to classes that were not used to learn the metrics. Using a metric learned on 1,000 classes, we show results for the ImageNet-10K dataset which contains 10,000 classes, and obtain performance that is competitive with the current state-of-the-art while being orders of magnitude faster. Furthermore, we show how a zero-shot class prior based on the ImageNet hierarchy can improve performance when few training images are available.

The second ‘chime’ speech separation and recognition challenge: Datasets, tasks and baselines

Abstract: Distant-microphone automatic speech recognition (ASR) remains a challenging goal in everyday environments involving multiple background sources and reverberation. This paper is intended to be a reference on the 2nd `CHiME' Challenge, an initiative designed to analyze and evaluate the performance of ASR systems in a real-world domestic environment. Two separate tracks have been proposed: a small-vocabulary task with small speaker movements and a medium-vocabulary task without speaker movements. We discuss the rationale for the challenge and provide a detailed description of the datasets, tasks and baseline performance results for each track.

MDPC-McEliece: New McEliece variants from Moderate Density Parity-Check codes

Abstract: In this work, we propose two McEliece variants: one from Moderate Density Parity-Check (MDPC) codes and another from quasi-cyclic MDPC codes. MDPC codes are LDPC codes of higher density (and worse error-correction capability) than what is usually adopted for telecommunication applications. However, in cryptography we are not necessarily interested in correcting many errors, but only a number which ensures an adequate security level. By this approach, we reduce under certain hypotheses the security of the scheme to the well studied decoding problem. Furthermore, the quasi-cyclic variant provides extremely compact-keys (for 80-bits of security, public-keys have only 4801 bits).

RIOT OS: Towards an OS for the Internet of Things

Abstract: The Internet of Things (IoT) is characterized by heterogeneous devices. They range from very lightweight sensors powered by 8-bit microcontrollers (MCUs) to devices equipped with more powerful, but energy-efficient 32-bit processors. Neither traditional operating systems (OS) currently running on Internet hosts, nor a typical OS for sensor networks are capable to fulfill the diverse requirements of such a wide range of devices. To leverage the IoT, redundant development should be avoided and maintenance costs should be reduced. In this paper we revisit the requirements for an OS in the IoT. We introduce RIOT OS, an OS that explicitly considers devices with minimal resources but eases development across a wide range of devices. RIOT OS allows for standard C and C++ programming, provides multi- threading as well as real-time capabilities, and needs only a minimum of 1.5 kB of RAM.

Depth synthesis and local warps for plausible image-based navigation

Abstract: Modern camera calibration and multiview stereo techniques enable users to smoothly navigate between different views of a scene captured using standard cameras. The underlying automatic 3D reconstruction methods work well for buildings and regular structures but often fail on vegetation, vehicles, and other complex geometry present in everyday urban scenes. Consequently, missing depth information makes Image-Based Rendering (IBR) for such scenes very challenging. Our goal is to provide plausible free-viewpoint navigation for such datasets. To do this, we introduce a new IBR algorithm that is robust to missing or unreliable geometry, providing plausible novel views even in regions quite far from the input camera positions. We first oversegment the input images, creating superpixels of homogeneous color content which often tends to preserve depth discontinuities. We then introduce a depth synthesis approach for poorly reconstructed regions based on a graph structure on the oversegmentation and appropriate traversal of the graph. The superpixels augmented with synthesized depth allow us to define a local shape-preserving warp which compensates for inaccurate depth. Our rendering algorithm blends the warped images, and generates plausible image-based novel views for our challenging target scenes. Our results demonstrate novel view synthesis in real time for multiple challenging scenes with significant depth complexity, providing a convincing immersive navigation experience.

Temperature effect on microalgae: a crucial factor for outdoor production

Abstract: High rate outdoor production units of microalgae can undergo temperature fluctuations. Seasonal temperature variations as well as more rapid daily fluctuations are liable to modify the growth conditions of microalgae and hence affect production efficiency. The effect of elevated temperatures, above optimal growth temperatures, on growth is seldom reported in literature, but often described as more deleterious than low temperatures. Depending on the species, different strategies are deployed to counteract the effect of above optimal temperatures such as energy re-balancing and cell shrinking. Moreover, long term adaptation of certain species over generation cycles has also been proven efficient to increase optimal temperatures. Physical models coupled to biological kinetics are able to predict the evolution of temperature in the growth media and its effect on the growth rate, highlighting the downstream drastic economic and environmental impacts. Regarding the relative elasticity of microalgae towards temperature issues, cell mortality can depend on species or adapted species and in certain cases can be attenuated. These elements can complement existing models and help visualize the effective impacts of temperature on outdoor cultures.

Learning-based approach for online lane change intention prediction

Abstract: Predicting driver behavior is a key component for Advanced Driver Assistance Systems (ADAS). In this paper, a novel approach based on Support Vector Machine and Bayesian filtering is proposed for online lane change intention prediction. The approach uses the multiclass probabilistic outputs of the Support Vector Machine as an input to the Bayesian filter, and the output of the Bayesian filter is used for the final prediction of lane changes. A lane tracker integrated in a passenger vehicle is used for real-world data collection for the purpose of training and testing. Data from different drivers on different highways were used to evaluate the robustness of the approach. The results demonstrate that the proposed approach is able to predict driver intention to change lanes on average 1.3 seconds in advance, with a maximum prediction horizon of 3.29 seconds.

CADP 2011: a toolbox for the construction and analysis of distributed processes

Abstract: CADP (Construction and Analysis of Distributed Processes) is a comprehensive software toolbox that implements the results of concurrency theory. Started in the mid 80s, CADP has been continuously developed by adding new tools and enhancing existing ones. Today, CADP benefits from a worldwide user community, both in academia and industry. This paper presents the latest release, CADP 2011, which is the result of a considerable development effort spanning the last five years. The paper first describes the theoretical principles and the modular architecture of CADP, which has inspired several other recent model checkers. The paper then reviews the main features of CADP 2011, including compilers for various formal specification languages, equivalence checkers, model checkers, compositional verification tools, performance evaluation tools, and parallel verification tools running on clusters and grids. Finally, the paper surveys some significant case studies.