Centre Inria de l'Université Grenoble Alpes

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.

This paper presents a method for recognizing scene categories based on approximate global geometric correspondence. This technique works by partitioning the image into increasingly fine sub-regions and computing histograms of local features found inside each sub-region. The resulting "spatial pyramid" is a simple and computationally efficient extension of an orderless bag-of-features image representation, and it shows significantly improved performance on challenging scene categorization tasks. Specifically, our proposed method exceeds the state of the art on the Caltech-101 database and achieves high accuracy on a large database of fifteen natural scene categories. The spatial pyramid framework also offers insights into the success of several recently proposed image descriptions, including Torralba’s "gist" and Lowe’s SIFT descriptors.

In this paper, we compare the performance of descriptors computed for local interest regions, as, for example, extracted by the Harris-Affine detector [Mikolajczyk, K and Schmid, C, 2004]. Many different descriptors have been proposed in the literature. It is unclear which descriptors are more appropriate and how their performance depends on the interest region detector. The descriptors should be distinctive and at the same time robust to changes in viewing conditions as well as to errors of the detector. Our evaluation uses as criterion recall with respect to precision and is carried out for different image transformations. We compare shape context [Belongie, S, et al., April 2002], steerable filters [Freeman, W and Adelson, E, Setp. 1991], PCA-SIFT [Ke, Y and Sukthankar, R, 2004], differential invariants [Koenderink, J and van Doorn, A, 1987], spin images [Lazebnik, S, et al., 2003], SIFT [Lowe, D. G., 1999], complex filters [Schaffalitzky, F and Zisserman, A, 2002], moment invariants [Van Gool, L, et al., 1996], and cross-correlation for different types of interest regions. We also propose an extension of the SIFT descriptor and show that it outperforms the original method. Furthermore, we observe that the ranking of the descriptors is mostly independent of the interest region detector and that the SIFT-based descriptors perform best. Moments and steerable filters show the best performance among the low dimensional descriptors.

In this paper, we question if self-supervised learning provides new properties to Vision Transformer (ViT) [16] that stand out compared to convolutional networks (convnets). Beyond the fact that adapting self-supervised methods to this architecture works particularly well, we make the following observations: first, self-supervised ViT features contain explicit information about the semantic segmentation of an image, which does not emerge as clearly with supervised ViTs, nor with convnets. Second, these features are also excellent k-NN classifiers, reaching 78.3% top-1 on ImageNet with a small ViT. Our study also underlines the importance of momentum encoder [26], multi-crop training [9], and the use of small patches with ViTs. We implement our findings into a simple self-supervised method, called DINO, which we interpret as a form of self-distillation with no labels. We show the synergy between DINO and ViTs by achieving 80.1% top-1 on ImageNet in linear evaluation with ViT-Base.

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.

We propose an assessing method of mixture model in a cluster analysis setting with integrated completed likelihood. For this purpose, the observed data are assigned to unknown clusters using a maximum a posteriori operator. Then, the integrated completed likelihood (ICL) is approximated using the Bayesian information criterion (BIC). Numerical experiments on simulated and real data of the resulting ICL criterion show that it performs well both for choosing a mixture model and a relevant number of clusters. In particular, ICL appears to be more robust than BIC to violation of some of the mixture model assumptions and it can select a number of dusters leading to a sensible partitioning of the data.

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.

This paper introduces a texture representation suitable for recognizing images of textured surfaces under a wide range of transformations, including viewpoint changes and nonrigid deformations. At the feature extraction stage, a sparse set of affine Harris and Laplacian regions is found in the image. Each of these regions can be thought of as a texture element having a characteristic elliptic shape and a distinctive appearance pattern. This pattern is captured in an affine-invariant fashion via a process of shape normalization followed by the computation of two novel descriptors, the spin image and the RIFT descriptor. When affine invariance is not required, the original elliptical shape servee as an additional discriminative feature for texture recognition. The proposed approach is evaluated in retrieval and classification tasks using the entire Brodatz database and a publicly available collection of 1,000 photographs of textured surfaces taken from different viewpoints.

This paper presents a new method for detecting scale invariant interest points. The method is based on two recent results on scale space: (1) Interest points can be adapted to scale and give repeatable results (geometrically stable). (2) Local extrema over scale of normalized derivatives indicate the presence of characteristic local structures. Our method first computes a multi-scale representation for the Harris interest point detector. We then select points at which a local measure (the Laplacian) is maximal over scales. This allows a selection of distinctive points for which the characteristic scale is known. These points are invariant to scale, rotation and translation as well as robust to illumination changes and limited changes of viewpoint. For indexing, the image is characterized by a set of scale invariant points; the scale associated with each point allows the computation of a scale invariant descriptor. Our descriptors are, in addition, invariant to image rotation, of affine illumination changes and robust to small perspective deformations. Experimental results for indexing show an excellent performance up to a scale factor of 4 for a database with more than 5000 images.

Abstract With the objective to improve road safety, the automotive industry is moving toward more “intelligent” vehicles. One of the major challenges is to detect dangerous situations and react accordingly in order to avoid or mitigate accidents. This requires predicting the likely evolution of the current traffic situation, and assessing how dangerous that future situation might be. This paper is a survey of existing methods for motion prediction and risk assessment for intelligent vehicles. The proposed classification is based on the semantics used to define motion and risk. We point out the tradeoff between model completeness and real-time constraints, and the fact that the choice of a risk assessment method is influenced by the selected motion model.

In this paper we compare the performance of interest point descriptors. Many different descriptors have been proposed in the literature. However, it is unclear which descriptors are more appropriate and how their performance depends on the interest point detector. The descriptors should be distinctive and at the same time robust to changes in viewing conditions as well as to errors of the point detector. Our evaluation uses as criterion detection rate with respect to false positive rate and is carried out for different image transformations. We compare SIFT descriptors (Lowe, 1999), steerable filters (Freeman and Adelson, 1991), differential invariants (Koenderink ad van Doorn, 1987), complex filters (Schaffalitzky and Zisserman, 2002), moment invariants (Van Gool et al., 1996) and cross-correlation for different types of interest points. In this evaluation, we observe that the ranking of the descriptors does not depend on the point detector and that SIFT descriptors perform best. Steerable filters come second ; they can be considered a good choice given the low dimensionality.

Typical human actions last several seconds and exhibit characteristic spatio-temporal structure. Recent methods attempt to capture this structure and learn action representations with convolutional neural networks. Such representations, however, are typically learned at the level of a few video frames failing to model actions at their full temporal extent. In this work we learn video representations using neural networks with long-term temporal convolutions (LTC). We demonstrate that LTC-CNN models with increased temporal extents improve the accuracy of action recognition. We also study the impact of different low-level representations, such as raw values of video pixels and optical flow vector fields and demonstrate the importance of high-quality optical flow estimation for learning accurate action models. We report state-of-the-art results on two challenging benchmarks for human action recognition UCF101 (92.7%) and HMDB51 (67.2%).

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.

Abstract On the basis of simulated data, this study compares the relative performances of the Bayesian clustering computer programs structure , geneland , geneclust and a new program named tess . While these four programs can detect population genetic structure from multilocus genotypes, only the last three ones include simultaneous analysis from geographical data. The programs are compared with respect to their abilities to infer the number of populations, to estimate membership probabilities, and to detect genetic discontinuities and clinal variation. The results suggest that combining analyses using tess and structure offers a convenient way to address inference of spatial population structure.

Attributes act as intermediate representations that enable 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 that 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. Label embedding enjoys a built-in ability to leverage alternative sources of information instead of or in addition to attributes, such as, e.g., class hierarchies or textual descriptions. Moreover, label embedding encompasses the whole range of learning settings from zero-shot learning to regular learning with a large number of labeled examples.

We propose a novel approach for optical flow estimation, targeted at large displacements with significant occlusions. It consists of two steps: i) dense matching by edge-preserving interpolation from a sparse set of matches; ii) variational energy minimization initialized with the dense matches. The sparse-to-dense interpolation relies on an appropriate choice of the distance, namely an edge-aware geodesic distance. This distance is tailored to handle occlusions and motion boundaries - two common and difficult issues for optical flow computation. We also propose an approximation scheme for the geodesic distance to allow fast computation without loss of performance. Subsequent to the dense interpolation step, standard one-level variational energy minimization is carried out on the dense matches to obtain the final flow estimation. The proposed approach, called Edge-Preserving Interpolation of Correspondences (EpicFlow) is fast and robust to large displacements. It significantly outperforms the state of the art on MPI-Sintel and performs on par on Kitti and Middlebury.

MOTIVATION: Polygenic scores have become a central tool in human genetics research. LDpred is a popular method for deriving polygenic scores based on summary statistics and a matrix of correlation between genetic variants. However, LDpred has limitations that may reduce its predictive performance. RESULTS: Here, we present LDpred2, a new version of LDpred that addresses these issues. We also provide two new options in LDpred2: a 'sparse' option that can learn effects that are exactly 0, and an 'auto' option that directly learns the two LDpred parameters from data. We benchmark predictive performance of LDpred2 against the previous version on simulated and real data, demonstrating substantial improvements in robustness and predictive accuracy compared to LDpred1. We then show that LDpred2 also outperforms other polygenic score methods recently developed, with a mean AUC over the 8 real traits analyzed here of 65.1%, compared to 63.8% for lassosum, 62.9% for PRS-CS and 61.5% for SBayesR. Note that LDpred2 provides more accurate polygenic scores when run genome-wide, instead of per chromosome. AVAILABILITY AND IMPLEMENTATION: LDpred2 is implemented in R package bigsnpr. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The high number of spectral bands acquired by hyperspectral sensors increases the capability to distinguish physical materials and objects, presenting new challenges to image analysis and classification. This letter presents a novel method for accurate spectral-spatial classification of hyperspectral images. The proposed technique consists of two steps. In the first step, a probabilistic support vector machine pixelwise classification of the hyperspectral image is applied. In the second step, spatial contextual information is used for refining the classification results obtained in the first step. This is achieved by means of a Markov random field regularization. Experimental results are presented for three hyperspectral airborne images and compared with those obtained by recently proposed advanced spectral-spatial classification techniques. The proposed method improves classification accuracies when compared to other classification approaches.

We describe a learning-based method for recovering 3D human body pose from single images and monocular image sequences. Our approach requires neither an explicit body model nor prior labeling of body parts in the image. Instead, it recovers pose by direct nonlinear regression against shape descriptor vectors extracted automatically from image silhouettes. For robustness against local silhouette segmentation errors, silhouette shape is encoded by histogram-of-shape-contexts descriptors. We evaluate several different regression methods: ridge regression, Relevance Vector Machine (RVM) regression, and Support Vector Machine (SVM) regression over both linear and kernel bases. The RVMs provide much sparser regressors without compromising performance, and kernel bases give a small but worthwhile improvement in performance. The loss of depth and limb labeling information often makes the recovery of 3D pose from single silhouettes ambiguous. To handle this, the method is embedded in a novel regressive tracking framework, using dynamics from the previous state estimate together with a learned regression value to disambiguate the pose. We show that the resulting system tracks long sequences stably. For realism and good generalization over a wide range of viewpoints, we train the regressors on images resynthesized from real human motion capture data. The method is demonstrated for several representations of full body pose, both quantitatively on independent but similar test data and qualitatively on real image sequences. Mean angular errors of 4-6 degrees are obtained for a variety of walking motions.

Color names are required in real-world applications such as image retrieval and image annotation. Traditionally, they are learned from a collection of labeled color chips. These color chips are labeled with color names within a well-defined experimental setup by human test subjects. However, naming colors in real-world images differs significantly from this experimental setting. In this paper, we investigate how color names learned from color chips compare to color names learned from real-world images. To avoid hand labeling real-world images with color names, we use Google Image to collect a data set. Due to the limitations of Google Image, this data set contains a substantial quantity of wrongly labeled data. We propose several variants of the PLSA model to learn color names from this noisy data. Experimental results show that color names learned from real-world images significantly outperform color names learned from labeled color chips for both image retrieval and image annotation.