Max Planck Institute for Informatics

Radiance Field methods have recently revolutionized novel-view synthesis of scenes captured with multiple photos or videos. However, achieving high visual quality still requires neural networks that are costly to train and render, while recent faster methods inevitably trade off speed for quality. For unbounded and complete scenes (rather than isolated objects) and 1080p resolution rendering, no current method can achieve real-time display rates. We introduce three key elements that allow us to achieve state-of-the-art visual quality while maintaining competitive training times and importantly allow high-quality real-time (≥ 30 fps) novel-view synthesis at 1080p resolution. First, starting from sparse points produced during camera calibration, we represent the scene with 3D Gaussians that preserve desirable properties of continuous volumetric radiance fields for scene optimization while avoiding unnecessary computation in empty space; Second, we perform interleaved optimization/density control of the 3D Gaussians, notably optimizing anisotropic covariance to achieve an accurate representation of the scene; Third, we develop a fast visibility-aware rendering algorithm that supports anisotropic splatting and both accelerates training and allows realtime rendering. We demonstrate state-of-the-art visual quality and real-time rendering on several established datasets.

While 3D object representations are being revived in the context of multi-view object class detection and scene understanding, they have not yet attained wide-spread use in fine-grained categorization. State-of-the-art approaches achieve remarkable performance when training data is plentiful, but they are typically tied to flat, 2D representations that model objects as a collection of unconnected views, limiting their ability to generalize across viewpoints. In this paper, we therefore lift two state-of-the-art 2D object representations to 3D, on the level of both local feature appearance and location. In extensive experiments on existing and newly proposed datasets, we show our 3D object representations outperform their state-of-the-art 2D counterparts for fine-grained categorization and demonstrate their efficacy for estimating 3D geometry from images via ultra-wide baseline matching and 3D reconstruction.

UNLABELLED: ROCR is a package for evaluating and visualizing the performance of scoring classifiers in the statistical language R. It features over 25 performance measures that can be freely combined to create two-dimensional performance curves. Standard methods for investigating trade-offs between specific performance measures are available within a uniform framework, including receiver operating characteristic (ROC) graphs, precision/recall plots, lift charts and cost curves. ROCR integrates tightly with R's powerful graphics capabilities, thus allowing for highly adjustable plots. Being equipped with only three commands and reasonable default values for optional parameters, ROCR combines flexibility with ease of usage. AVAILABILITY: http://rocr.bioinf.mpi-sb.mpg.de. ROCR can be used under the terms of the GNU General Public License. Running within R, it is platform-independent. CONTACT: tobias.sing@mpi-sb.mpg.de.

Pedestrian detection is a key problem in computer vision, with several applications that have the potential to positively impact quality of life. In recent years, the number of approaches to detecting pedestrians in monocular images has grown steadily. However, multiple data sets and widely varying evaluation protocols are used, making direct comparisons difficult. To address these shortcomings, we perform an extensive evaluation of the state of the art in a unified framework. We make three primary contributions: 1) We put together a large, well-annotated, and realistic monocular pedestrian detection data set and study the statistics of the size, position, and occlusion patterns of pedestrians in urban scenes, 2) we propose a refined per-frame evaluation methodology that allows us to carry out probing and informative comparisons, including measuring performance in relation to scale and occlusion, and 3) we evaluate the performance of sixteen pretrained state-of-the-art detectors across six data sets. Our study allows us to assess the state of the art and provides a framework for gauging future efforts. Our experiments show that despite significant progress, performance still has much room for improvement. In particular, detection is disappointing at low resolutions and for partially occluded pedestrians.

Human pose estimation has made significant progress during the last years. However current datasets are limited in their coverage of the overall pose estimation challenges. Still these serve as the common sources to evaluate, train and compare different models on. In this paper we introduce a novel benchmark "MPII Human Pose" that makes a significant advance in terms of diversity and difficulty, a contribution that we feel is required for future developments in human body models. This comprehensive dataset was collected using an established taxonomy of over 800 human activities [1]. The collected images cover a wider variety of human activities than previous datasets including various recreational, occupational and householding activities, and capture people from a wider range of viewpoints. We provide a rich set of labels including positions of body joints, full 3D torso and head orientation, occlusion labels for joints and body parts, and activity labels. For each image we provide adjacent video frames to facilitate the use of motion information. Given these rich annotations we perform a detailed analysis of leading human pose estimation approaches and gaining insights for the success and failures of these methods.

MOTIVATION: In life sciences, interpretability of machine learning models is as important as their prediction accuracy. Linear models are probably the most frequently used methods for assessing feature relevance, despite their relative inflexibility. However, in the past years effective estimators of feature relevance have been derived for highly complex or non-parametric models such as support vector machines and RandomForest (RF) models. Recently, it has been observed that RF models are biased in such a way that categorical variables with a large number of categories are preferred. RESULTS: In this work, we introduce a heuristic for normalizing feature importance measures that can correct the feature importance bias. The method is based on repeated permutations of the outcome vector for estimating the distribution of measured importance for each variable in a non-informative setting. The P-value of the observed importance provides a corrected measure of feature importance. We apply our method to simulated data and demonstrate that (i) non-informative predictors do not receive significant P-values, (ii) informative variables can successfully be recovered among non-informative variables and (iii) P-values computed with permutation importance (PIMP) are very helpful for deciding the significance of variables, and therefore improve model interpretability. Furthermore, PIMP was used to correct RF-based importance measures for two real-world case studies. We propose an improved RF model that uses the significant variables with respect to the PIMP measure and show that its prediction accuracy is superior to that of other existing models. AVAILABILITY: R code for the method presented in this article is available at http://www.mpi-inf.mpg.de/ approximately altmann/download/PIMP.R CONTACT: altmann@mpi-inf.mpg.de, laura.tolosi@mpi-inf.mpg.de SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

This paper provides an entry point to the problem of interpreting a deep neural network model and explaining its predictions. It is based on a tutorial given at ICASSP 2017. As a tutorial paper, the set of methods covered here is not exhaustive, but sufficiently representative to discuss a number of questions in interpretability, technical challenges, and possible applications. The second part of the tutorial focuses on the recently proposed layer-wise relevance propagation (LRP) technique, for which we provide theory, recommendations, and tricks, to make most efficient use of it on real data.

Article Share on Parallel & distributed processing Authors: Philipp Slusallek Saarland University, Saarbrücken, Germany Saarland University, Saarbrücken, GermanyView Profile , Peter Shirley University of Utah, Salt Lake City, UT University of Utah, Salt Lake City, UTView Profile , William Mark University of Texas at Austin, Austin, TX University of Texas at Austin, Austin, TXView Profile , Gordon Stoll Intel Corporation, Santa Clara, CA Intel Corporation, Santa Clara, CAView Profile , Ingo Wald Max-Planck-Institut für Informatik, Saarbrücken, Germany Max-Planck-Institut für Informatik, Saarbrücken, GermanyView Profile Authors Info & Claims SIGGRAPH '05: ACM SIGGRAPH 2005 CoursesJuly 2005 Pages 11–eshttps://doi.org/10.1145/1198555.1198750Published:31 July 2005Publication History 1citation305DownloadsMetricsTotal Citations1Total Downloads305Last 12 Months5Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

MOTIVATION: The result of a typical microarray experiment is a long list of genes with corresponding expression measurements. This list is only the starting point for a meaningful biological interpretation. Modern methods identify relevant biological processes or functions from gene expression data by scoring the statistical significance of predefined functional gene groups, e.g. based on Gene Ontology (GO). We develop methods that increase the explanatory power of this approach by integrating knowledge about relationships between the GO terms into the calculation of the statistical significance. RESULTS: We present two novel algorithms that improve GO group scoring using the underlying GO graph topology. The algorithms are evaluated on real and simulated gene expression data. We show that both methods eliminate local dependencies between GO terms and point to relevant areas in the GO graph that remain undetected with state-of-the-art algorithms for scoring functional terms. A simulation study demonstrates that the new methods exhibit a higher level of detecting relevant biological terms than competing methods.

Deep learning has led to a paradigm shift in artificial intelligence, including web, text, and image search, speech recognition, as well as bioinformatics, with growing impact in chemical physics. Machine learning, in general, and deep learning, in particular, are ideally suitable for representing quantum-mechanical interactions, enabling us to model nonlinear potential-energy surfaces or enhancing the exploration of chemical compound space. Here we present the deep learning architecture SchNet that is specifically designed to model atomistic systems by making use of continuous-filter convolutional layers. We demonstrate the capabilities of SchNet by accurately predicting a range of properties across chemical space for molecules and materials, where our model learns chemically plausible embeddings of atom types across the periodic table. Finally, we employ SchNet to predict potential-energy surfaces and energy-conserving force fields for molecular dynamics simulations of small molecules and perform an exemplary study on the quantum-mechanical properties of C20-fullerene that would have been infeasible with regular ab initio molecular dynamics.

This paper presents a method for face recognition across variations in pose, ranging from frontal to profile views, and across a wide range of illuminations, including cast shadows and specular reflections. To account for these variations, the algorithm simulates the process of image formation in 3D space, using computer graphics, and it estimates 3D shape and texture of faces from single images. The estimate is achieved by fitting a statistical, morphable model of 3D faces to images. The model is learned from a set of textured 3D scans of heads. We describe the construction of the morphable model, an algorithm to fit the model to images, and a framework for face identification. In this framework, faces are represented by model parameters for 3D shape and texture. We present results obtained with 4,488 images from the publicly available CMU-PIE database and 1,940 images from the FERET database.

UNLABELLED: Rapidly increasing amounts of molecular interaction data are being produced by various experimental techniques and computational prediction methods. In order to gain insight into the organization and structure of the resultant large complex networks formed by the interacting molecules, we have developed the versatile Cytoscape plugin NetworkAnalyzer. It computes and displays a comprehensive set of topological parameters, which includes the number of nodes, edges, and connected components, the network diameter, radius, density, centralization, heterogeneity, and clustering coefficient, the characteristic path length, and the distributions of node degrees, neighborhood connectivities, average clustering coefficients, and shortest path lengths. NetworkAnalyzer can be applied to both directed and undirected networks and also contains extra functionality to construct the intersection or union of two networks. It is an interactive and highly customizable application that requires no expert knowledge in graph theory from the user. AVAILABILITY: NetworkAnalyzer can be downloaded via the Cytoscape web site: http://www.cytoscape.org

We present a novel approach for real-time facial reenactment of a monocular target video sequence (e.g., Youtube video). The source sequence is also a monocular video stream, captured live with a commodity webcam. Our goal is to animate the facial expressions of the target video by a source actor and re-render the manipulated output video in a photo-realistic fashion. To this end, we first address the under-constrained problem of facial identity recovery from monocular video by non-rigid model-based bundling. At run time, we track facial expressions of both source and target video using a dense photometric consistency measure. Reenactment is then achieved by fast and efficient deformation transfer between source and target. The mouth interior that best matches the re-targeted expression is retrieved from the target sequence and warped to produce an accurate fit. Finally, we convincingly re-render the synthesized target face on top of the corresponding video stream such that it seamlessly blends with the real-world illumination. We demonstrate our method in a live setup, where Youtube videos are reenacted in real time.

The last 20 years have seen ever-increasing research activity in the field of human activity recognition. With activity recognition having considerably matured, so has the number of challenges in designing, implementing, and evaluating activity recognition systems. This tutorial aims to provide a comprehensive hands-on introduction for newcomers to the field of human activity recognition. It specifically focuses on activity recognition using on-body inertial sensors. We first discuss the key research challenges that human activity recognition shares with general pattern recognition and identify those challenges that are specific to human activity recognition. We then describe the concept of an Activity Recognition Chain (ARC) as a general-purpose framework for designing and evaluating activity recognition systems. We detail each component of the framework, provide references to related research, and introduce the best practice methods developed by the activity recognition research community. We conclude with the educational example problem of recognizing different hand gestures from inertial sensors attached to the upper and lower arm. We illustrate how each component of this framework can be implemented for this specific activity recognition problem and demonstrate how different implementations compare and how they impact overall recognition performance.

Due to the importance of zero-shot learning, i.e., classifying images where there is a lack of labeled training data, the number of proposed approaches has recently increased steadily. We argue that it is time to take a step back and to analyze the status quo of the area. The purpose of this paper is three-fold. First, given the fact that there is no agreed upon zero-shot learning benchmark, we first define a new benchmark by unifying both the evaluation protocols and data splits of publicly available datasets used for this task. This is an important contribution as published results are often not comparable and sometimes even flawed due to, e.g., pre-training on zero-shot test classes. Moreover, we propose a new zero-shot learning dataset, the Animals with Attributes 2 (AWA2) dataset which we make publicly available both in terms of image features and the images themselves. Second, we compare and analyze a significant number of the state-of-the-art methods in depth, both in the classic zero-shot setting but also in the more realistic generalized zero-shot setting. Finally, we discuss in detail the limitations of the current status of the area which can be taken as a basis for advancing it.

Automatic synthesis of realistic images from text would be interesting and useful, but current AI systems are still far from this goal. However, in recent years generic and powerful recurrent neural network architectures have been developed to learn discriminative text feature representations. Meanwhile, deep convolutional generative adversarial networks (GANs) have begun to generate highly compelling images of specific categories, such as faces, album covers, and room interiors. In this work, we develop a novel deep architecture and GAN formulation to effectively bridge these advances in text and image model- ing, translating visual concepts from characters to pixels. We demonstrate the capability of our model to generate plausible images of birds and flowers from detailed text descriptions.

The recent boom in microfluidics and combinatorial indexing strategies, combined with low sequencing costs, has empowered single-cell sequencing technology. Thousands-or even millions-of cells analyzed in a single experiment amount to a data revolution in single-cell biology and pose unique data science problems. Here, we outline eleven challenges that will be central to bringing this emerging field of single-cell data science forward. For each challenge, we highlight motivating research questions, review prior work, and formulate open problems. This compendium is for established researchers, newcomers, and students alike, highlighting interesting and rewarding problems for the coming years.

Modeling textual or visual information with vector representations trained from large language or visual datasets has been successfully explored in recent years. However, tasks such as visual question answering require combining these vector representations with each other. Approaches to multimodal pooling include element-wise product or sum, as well as concatenation of the visual and textual representations. We hypothesize that these methods are not as expressive as an outer product of the visual and textual vectors. As the outer product is typically infeasible due to its high dimensionality, we instead propose utilizing Multimodal Compact Bilinear pooling (MCB) to efficiently and expressively combine multimodal features. We extensively evaluate MCB on the visual question answering and grounding tasks. We consistently show the benefit of MCB over ablations without MCB. For visual question answering, we present an architecture which uses MCB twice, once for predicting attention over spatial features and again to combine the attended representation with the question representation. This model outperforms the state-of-the-art on the Visual7W dataset and the VQA challenge.

Meta-learning has been proposed as a framework to address the challenging few-shot learning setting. The key idea is to leverage a large number of similar few-shot tasks in order to learn how to adapt a base-learner to a new task for which only a few labeled samples are available. As deep neural networks (DNNs) tend to overfit using a few samples only, meta-learning typically uses shallow neural networks (SNNs), thus limiting its effectiveness. In this paper we propose a novel few-shot learning method called meta-transfer learning (MTL) which learns to adapt a deep NN for few shot learning tasks. Specifically, "meta" refers to training multiple tasks, and "transfer" is achieved by learning scaling and shifting functions of DNN weights for each task. In addition, we introduce the hard task (HT) meta-batch scheme as an effective learning curriculum for MTL. We conduct experiments using (5-class, 1-shot) and (5-class, 5-shot) recognition tasks on two challenging few-shot learning benchmarks: miniImageNet and Fewshot-CIFAR100. Extensive comparisons to related works validate that our meta-transfer learning approach trained with the proposed HT meta-batch scheme achieves top performance. An ablation study also shows that both components contribute to fast convergence and high accuracy.

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works