Sebastian Nowozin

On feature combination for multiclass object classification

A key ingredient in the design of visual object classification systems is the identification of relevant class specific aspects while being robust to intra-class variations. While this is a necessity in order to generalize beyond a given set of training images, it is also a very difficult problem due to the high variability of visual appearance within each class. In the last years substantial performance gains on challenging benchmark datasets have been reported in the literature. This progress can be attributed to two developments: the design of highly discriminative and robust image features and the combination of multiple complementary features based on different aspects such as shape, color or texture. In this paper we study several models that aim at learning the correct weighting of different features from training data. These include multiple kernel learning as well as simple baseline methods. Furthermore we derive ensemble methods inspired by Boosting which are easily extendable to several multiclass setting. All methods are thoroughly evaluated on object classification datasets using a multitude of feature descriptors. The key results are that even very simple baseline methods, that are orders of magnitude faster than learning techniques are highly competitive with multiple kernel learning. Furthermore the Boosting type methods are found to produce consistently better results in all experiments. We provide insight of when combination methods can be expected to work and how the benefit of complementary features can be exploited most efficiently.

Instructing people for training gestural interactive systems

Entertainment and gaming systems such as the Wii and XBox Kinect have brought touchless, body-movement based interfaces to the masses. Systems like these enable the estimation of movements of various body parts from raw inertial motion or depth sensor data. However, the interface developer is still left with the challenging task of creating a system that recognizes these movements as embodying meaning. The machine learning approach for tackling this problem requires the collection of data sets that contain the relevant body movements and their associated semantic labels. These data sets directly impact the accuracy and performance of the gesture recognition system and should ideally contain all natural variations of the movements associated with a gesture. This paper addresses the problem of collecting such gesture datasets. In particular, we investigate the question of what is the most appropriate semiotic modality of instructions for conveying to human subjects the movements the system developer needs them to perform. The results of our qualitative and quantitative analysis indicate that the choice of modality has a significant impact on the performance of the learnt gesture recognition system; particularly in terms of correctness and coverage.

Structured Learning and Prediction in Computer Vision

Powerful statistical models that can be learned efficiently from large amounts of data are currently revolutionizing computer vision. These models possess a rich internal structure reflecting task-specific relations and constraints. This monograph introduces the reader to the most popular classes of structured models in computer vision. Our focus is discrete undirected graphical models which we cover in detail together with a description of algorithms for both probabilistic inference and maximum a posteriori inference. We discuss separately recently successful techniques for prediction in general structured models. In the second part of this monograph we describe methods for parameter learning where we distinguish the classic maximum likelihood based methods from the more recent prediction-based parameter learning methods. We highlight developments to enhance current models and discuss kernelized models and latent variable models. To make the monograph more practical and to provide links to further study we provide examples of successful application of many methods in the computer vision literature.

Discriminative Subsequence Mining for Action Classification

Recent approaches to action classification in videos have used sparse spatio-temporal words encoding local appearance around interesting movements. Most of these approaches use a histogram representation, discarding the temporal order among features. But this ordering information can contain important information about the action itself e.g. consider the sport disciplines of hurdle race and long jump, where the global temporal order of motions (running, jumping) is important to discriminate between the two. In this work we propose to use a sequential representation which retains this temporal order. Further, we introduce Discriminative Subsequence Mining to find optimal discriminative subsequence patterns. In combination with the LPBoost classifier, this amounts to simultaneously learning a classification function and performing feature selection in the space of all possible feature sequences. The resulting classifier linearly combines a small number of interpretable decision functions, each checking for the presence of a single discriminative pattern. The classifier is benchmarked on the KTH action classification data set and outperforms the best known results in the literature.

A Comparative Study of Modern Inference Techniques for Discrete Energy Minimization Problems

Even years ago, Szeliski et al. published an influential study on energy minimization methods for Markov random fields (MRF). This study provided valuable insights in choosing the best optimization technique for certain classes of problems. While these insights remain generally useful today, the phenominal success of random field models means that the kinds of inference problems we solve have changed significantly. Specifically, the models today often include higher order interactions, flexible connectivity structures, large label-spaces of different cardinalities, or learned energy tables. To reflect these changes, we provide a modernized and enlarged study. We present an empirical comparison of 24 state-of-art techniques on a corpus of 2,300 energy minimization instances from 20 diverse computer vision applications. To ensure reproducibility, we evaluate all methods in the OpenGM2 framework and report extensive results regarding runtime and solution quality. Key insights from our study agree with the results of Szeliski et al. for the types of models they studied. However, on new and challenging types of models our findings disagree and suggest that polyhedral methods and integer programming solvers are competitive in terms of runtime and solution quality over a large range of model types.

gBoost: a mathematical programming approach to graph classification and regression

Graph mining methods enumerate frequently appearing subgraph patterns, which can be used as features for subsequent classification or regression. However, frequent patterns are not necessarily informative for the given learning problem. We propose a mathematical programming boosting method (gBoost) that progressively collects informative patterns. Compared to AdaBoost, gBoost can build the prediction rule with fewer iterations. To apply the boosting method to graph data, a branch-and-bound pattern search algorithm is developed based on the DFS code tree. The constructed search space is reused in later iterations to minimize the computation time. Our method can learn more efficiently than the simpler method based on frequent substructure mining, because the output labels are used as an extra information source for pruning the search space. Furthermore, by engineering the mathematical program, a wide range of machine learning problems can be solved without modifying the pattern search algorithm.

Decision tree fields

This paper introduces a new formulation for discrete image labeling tasks, the Decision Tree Field (DTF), that combines and generalizes random forests and conditional random fields (CRF) which have been widely used in computer vision. In a typical CRF model the unary potentials are derived from sophisticated random forest or boosting based classifiers, however, the pairwise potentials are assumed to (1) have a simple parametric form with a pre-specified and fixed dependence on the image data, and (2) to be defined on the basis of a small and fixed neighborhood. In contrast, in DTF, local interactions between multiple variables are determined by means of decision trees evaluated on the image data, allowing the interactions to be adapted to the image content. This results in powerful graphical models which are able to represent complex label structure. Our key technical contribution is to show that the DTF model can be trained efficiently and jointly using a convex approximate likelihood function, enabling us to learn over a million free model parameters. We show experimentally that for applications which have a rich and complex label structure, our model achieves excellent results.

A Comparative Study of Modern Inference Techniques for Structured Discrete Energy Minimization Problems

Szeliski et al. published an influential study in 2006 on energy minimization methods for Markov random fields. This study provided valuable insights in choosing the best optimization technique for certain classes of problems. While these insights remain generally useful today, the phenomenal success of random field models means that the kinds of inference problems that have to be solved changed significantly. Specifically, the models today often include higher order interactions, flexible connectivity structures, large label-spaces of different cardinalities, or learned energy tables. To reflect these changes, we provide a modernized and enlarged study. We present an empirical comparison of more than 27 state-of-the-art optimization techniques on a corpus of 2453 energy minimization instances from diverse applications in computer vision. To ensure reproducibility, we evaluate all methods in the OpenGM 2 framework and report extensive results regarding runtime and solution quality. Key insights from our study agree with the results of Szeliski et al. for the types of models they studied. However, on new and challenging types of models our findings disagree and suggest that polyhedral methods and integer programming solvers are competitive in terms of runtime and solution quality over a large range of model types.

Weighted Substructure Mining for Image Analysis

In Web-related applications of image categorization, it is desirable to derive an interpretable classification rule with high accuracy. Using the bag-of-words representation and the linear support vector machine, one can partly fulfill the goal, but the accuracy of linear classifiers is not high and the obtained features are not informative for users. We propose to combine item set mining and large margin classifiers to select features from the power set of all visual words. Our resulting classification rule is easier to browse and simpler to understand, because each feature has richer information. As a next step, each image is represented as a graph where nodes correspond to local image features and edges encode geometric relations between features. Combining graph mining and boosting, we can obtain a classification rule based on subgraph features that contain more information than the set features. We evaluate our algorithm in a web-retrieval ranking task where the goal is to reject outliers from a set of images returned for a keyword query. Furthermore, it is evaluated on the supervised classification tasks with the challenging VOC2005 data set. Our approach yields excellent accuracy in the unsupervised ranking task compared to a recently proposed probabilistic model and competitive results in the supervised classification task.

On parameter learning in CRF-based approaches to object class image segmentation

Recent progress in per-pixel object class labeling of natural images can be attributed to the use of multiple types of image features and sound statistical learning approaches. Within the latter, Conditional Random Fields (CRF) are prominently used for their ability to represent interactions between random variables. Despite their popularity in computer vision, parameter learning for CRFs has remained difficult, popular approaches being cross-validation and piecewise training. In this work, we propose a simple yet expressive tree-structured CRF based on a recent hierarchical image segmentation method. Our model combines and weights multiple image features within a hierarchical representation and allows simple and efficient globally-optimal learning of ≅ 105 parameters. The tractability of our model allows us to pose and answer some of the open questions regarding parameter learning applying to CRF-based approaches. The key findings for learning CRF models are, from the obvious to the surprising, i) multiple image features always help, ii) the limiting dimension with respect to current models is the amount of training data, iii) piecewise training is competitive, iv) current methods for max-margin training fail for models with many parameters.