Hanno Ackermann

Risto Vuorio, Pim de Haan, Johann Brehmer et al.

Standard imitation learning can fail when the expert demonstrators have different sensory inputs than the imitating agent. This is because partial observability gives rise to hidden confounders in the causal graph. In previous work, to work around the confounding problem, policies have been trained using query access to the expert's policy or inverse reinforcement learning (IRL). However, both approaches have drawbacks as the expert's policy may not be available and IRL can be unstable in practice. Instead, we propose to train a variational inference model to infer the expert's latent information and use it to train a latent-conditional policy. We prove that using this method, under strong assumptions, the identification of the correct imitation learning policy is theoretically possible from expert demonstrations alone. In practice, we focus on a setting with less strong assumptions where we use exploration data for learning the inference model. We show in theory and practice that this algorithm converges to the correct interventional policy, solves the confounding issue, and can under certain assumptions achieve an asymptotically optimal imitation performance.

Hanno Ackermann, Hong Cai, Mohsen Ghafoorian et al.

The attention mechanism is an important reason for the success of transformers. It relies on computing pairwise relations between tokens. To reduce the high computational cost of standard quadratic attention, linear attention has been proposed as an efficient approximation. It employs kernel functions that are applied independently to the inputs before the pairwise similarities are calculated. That allows for an efficient computational procedure which, however, amounts to a low-degree rational function approximating softmax. We propose Hadamard Linear Attention (HLA). Unlike previous works on linear attention, the nonlinearity in HLA is not applied separately to queries and keys, but, analogously to standard softmax attention, after the pairwise similarities have been computed. It will be shown that the proposed nonlinearity amounts to a higher-degree rational function to approximate softmax. An efficient computational scheme for the proposed method is derived that is similar to that of standard linear attention. In contrast to other approaches, no time-consuming tensor reshaping is necessary to apply the proposed algorithm. The effectiveness of the approach is demonstrated by applying it to a large diffusion transformer model for video generation, an application that involves very large amounts of tokens.

Rajeev Yasarla, Deepti Hegde, Hsin-Pai Cheng et al.

Vision-language-action (VLA) models are effective as end-to-end motion planners, but can be brittle when evaluated in closed-loop settings due to being trained under traditional imitation learning framework. Existing closed-loop supervision approaches lack scalability and fail to completely model a reactive environment. We propose MAPLE, a novel framework for reactive, multi-agent rollout of a dynamic driving scenario in the latent space of the VLA model. The ego vehicle and nearby traffic agents are independently controlled over multi-step horizons, while being reactive to other agents in the scene, enabling closed-loop training. MAPLE consists of two training stages: (1) supervised fine-tuning on the latent rollouts based on ground-truth trajectories, followed by (2) reinforcement learning with global and agent -specific rewards that encourage safety, progress, and interaction realism. We further propose diversity rewards that encourage the model to generate planning behaviors that may not be present in logged driving data. Notably, our closed-loop training framework is scalable and does not require external simulators, which can be computationally expensive to run and have limited visual fidelity to the real-world. MAPLE achieves state-of-the-art driving performance on Bench2Drive and demonstrates scalable, closed-loop multi-agent play for robust E2E autonomous driving systems.

Yuren Cong, Wentong Liao, Hanno Ackermann et al.

Dynamic scene graph generation aims at generating a scene graph of the given video. Compared to the task of scene graph generation from images, it is more challenging because of the dynamic relationships between objects and the temporal dependencies between frames allowing for a richer semantic interpretation. In this paper, we propose Spatial-temporal Transformer (STTran), a neural network that consists of two core modules: (1) a spatial encoder that takes an input frame to extract spatial context and reason about the visual relationships within a frame, and (2) a temporal decoder which takes the output of the spatial encoder as input in order to capture the temporal dependencies between frames and infer the dynamic relationships. Furthermore, STTran is flexible to take varying lengths of videos as input without clipping, which is especially important for long videos. Our method is validated on the benchmark dataset Action Genome (AG). The experimental results demonstrate the superior performance of our method in terms of dynamic scene graphs. Moreover, a set of ablative studies is conducted and the effect of each proposed module is justified. Code available at: https://github.com/yrcong/STTran.

Farhad G. Zanjani, Hong Cai, Hanno Ackermann et al.

This paper presents Planar Gaussian Splatting (PGS), a novel neural rendering approach to learn the 3D geometry and parse the 3D planes of a scene, directly from multiple RGB images. The PGS leverages Gaussian primitives to model the scene and employ a hierarchical Gaussian mixture approach to group them. Similar Gaussians are progressively merged probabilistically in the tree-structured Gaussian mixtures to identify distinct 3D plane instances and form the overall 3D scene geometry. In order to enable the grouping, the Gaussian primitives contain additional parameters, such as plane descriptors derived by lifting 2D masks from a general 2D segmentation model and surface normals. Experiments show that the proposed PGS achieves state-of-the-art performance in 3D planar reconstruction without requiring either 3D plane labels or depth supervision. In contrast to existing supervised methods that have limited generalizability and struggle under domain shift, PGS maintains its performance across datasets thanks to its neural rendering and scene-specific optimization mechanism, while also being significantly faster than existing optimization-based approaches.

Evgenii Egorov, Hanno Ackermann, Markus Nagel et al.

Attention layers apply a sequence-to-sequence mapping whose parameters depend on the pairwise interactions of the input elements. However, without any structural assumptions, memory and compute scale quadratically with the sequence length. The two main ways to mitigate this are to introduce sparsity by ignoring a sufficient amount of pairwise interactions or to introduce recurrent dependence along them, as SSM does. Although both approaches are reasonable, they both have disadvantages. We propose a novel algorithm that combines the advantages of both concepts. Our idea is based on the efficient inversion of tree-structured matrices.

Leyla Mirvakhabova, Hong Cai, Jisoo Jeong et al.

Recent works on optical flow estimation use neural networks to predict the flow field that maps positions of one image to positions of the other. These networks consist of a feature extractor, a correlation volume, and finally several refinement steps. These refinement steps mimic the iterative refinements performed by classical optimization algorithms and are usually implemented by neural layers (e.g., GRU) which are recurrently executed for a fixed and pre-determined number of steps. However, relying on a fixed number of steps may result in suboptimal performance because it is not tailored to the input data. In this paper, we introduce a novel approach for predicting the derivative of the flow using a continuous model, namely neural ordinary differential equations (ODE). One key advantage of this approach is its capacity to model an equilibrium process, dynamically adjusting the number of compute steps based on the data at hand. By following a particular neural architecture, ODE solver, and associated hyperparameters, our proposed model can replicate the exact same updates as recurrent cells used in existing works, offering greater generality. Through extensive experimental analysis on optical flow benchmarks, we demonstrate that our approach achieves an impressive improvement over baseline and existing models, all while requiring only a single refinement step.

Florian Kluger, Hanno Ackermann, Eric Brachmann et al.

Humans perceive and construct the surrounding world as an arrangement of simple parametric models. In particular, man-made environments commonly consist of volumetric primitives such as cuboids or cylinders. Inferring these primitives is an important step to attain high-level, abstract scene descriptions. Previous approaches directly estimate shape parameters from a 2D or 3D input, and are only able to reproduce simple objects, yet unable to accurately parse more complex 3D scenes. In contrast, we propose a robust estimator for primitive fitting, which can meaningfully abstract real-world environments using cuboids. A RANSAC estimator guided by a neural network fits these primitives to 3D features, such as a depth map. We condition the network on previously detected parts of the scene, thus parsing it one-by-one. To obtain 3D features from a single RGB image, we additionally optimise a feature extraction CNN in an end-to-end manner. However, naively minimising point-to-primitive distances leads to large or spurious cuboids occluding parts of the scene behind. We thus propose an occlusion-aware distance metric correctly handling opaque scenes. The proposed algorithm does not require labour-intensive labels, such as cuboid annotations, for training. Results on the challenging NYU Depth v2 dataset demonstrate that the proposed algorithm successfully abstracts cluttered real-world 3D scene layouts.

Cong Yuren, Hanno Ackermann, Wentong Liao et al.

Semantic image understanding is a challenging topic in computer vision. It requires to detect all objects in an image, but also to identify all the relations between them. Detected objects, their labels and the discovered relations can be used to construct a scene graph which provides an abstract semantic interpretation of an image. In previous works, relations were identified by solving an assignment problem formulated as Mixed-Integer Linear Programs. In this work, we interpret that formulation as Ordinary Differential Equation (ODE). The proposed architecture performs scene graph inference by solving a neural variant of an ODE by end-to-end learning. It achieves state-of-the-art results on all three benchmark tasks: scene graph generation (SGGen), classification (SGCls) and visual relationship detection (PredCls) on Visual Genome benchmark.

Florian Kluger, Eric Brachmann, Hanno Ackermann et al.

We present a robust estimator for fitting multiple parametric models of the same form to noisy measurements. Applications include finding multiple vanishing points in man-made scenes, fitting planes to architectural imagery, or estimating multiple rigid motions within the same sequence. In contrast to previous works, which resorted to hand-crafted search strategies for multiple model detection, we learn the search strategy from data. A neural network conditioned on previously detected models guides a RANSAC estimator to different subsets of all measurements, thereby finding model instances one after another. We train our method supervised as well as self-supervised. For supervised training of the search strategy, we contribute a new dataset for vanishing point estimation. Leveraging this dataset, the proposed algorithm is superior with respect to other robust estimators as well as to designated vanishing point estimation algorithms. For self-supervised learning of the search, we evaluate the proposed algorithm on multi-homography estimation and demonstrate an accuracy that is superior to state-of-the-art methods.

Maren Awiszus, Hanno Ackermann, Bodo Rosenhahn

Image generating neural networks are mostly viewed as black boxes, where any change in the input can have a number of globally effective changes on the output. In this work, we propose a method for learning disentangled representations to allow for localized image manipulations. We use face images as our example of choice. Depending on the image region, identity and other facial attributes can be modified. The proposed network can transfer parts of a face such as shape and color of eyes, hair, mouth, etc.~directly between persons while all other parts of the face remain unchanged. The network allows to generate modified images which appear like realistic images. Our model learns disentangled representations by weak supervision. We propose a localized resnet autoencoder optimized using several loss functions including a loss based on the semantic segmentation, which we interpret as masks, and a loss which enforces disentanglement by decomposition of the latent space into statistically independent subspaces. We evaluate the proposed solution w.r.t. disentanglement and generated image quality. Convincing results are demonstrated using the CelebA dataset.

Florian Kluger, Hanno Ackermann, Michael Ying Yang et al.

The horizon line is an important geometric feature for many image processing and scene understanding tasks in computer vision. For instance, in navigation of autonomous vehicles or driver assistance, it can be used to improve 3D reconstruction as well as for semantic interpretation of dynamic environments. While both algorithms and datasets exist for single images, the problem of horizon line estimation from video sequences has not gained attention. In this paper, we show how convolutional neural networks are able to utilise the temporal consistency imposed by video sequences in order to increase the accuracy and reduce the variance of horizon line estimates. A novel CNN architecture with an improved residual convolutional LSTM is presented for temporally consistent horizon line estimation. We propose an adaptive loss function that ensures stable training as well as accurate results. Furthermore, we introduce an extension of the KITTI dataset which contains precise horizon line labels for 43699 images across 72 video sequences. A comprehensive evaluation shows that the proposed approach consistently achieves superior performance compared with existing methods.

Sami S. Brandt, Hanno Ackermann

In this paper, we show that the affine, non-rigid structure-from-motion problem can be solved by rank-one, thus degenerate, basis shapes. It is a natural reformulation of the classic low-rank method by Bregler et al., where it was assumed that the deformable 3D structure is generated by a linear combination of rigid basis shapes. The non-rigid shape will be decomposed into the mean shape and the degenerate shapes, constructed from the right singular vectors of the low-rank decomposition. The right singular vectors are affinely back-projected into the 3D space, and the affine back-projections will also be solved as part of the factorisation. By construction, a direct interpretation for the right singular vectors of the low-rank decomposition will also follow: they can be seen as principal components, hence, the first variant of our method is referred to as Rank-1-PCA. The second variant, referred to as Rank-1-ICA, additionally estimates the orthogonal transform which maps the deformation modes into as statistically independent modes as possible. It has the advantage of pinpointing statistically dependent subspaces related to, for instance, lip movements on human faces. Moreover, in contrast to prior works, no predefined dimensionality for the subspaces is imposed. The experiments on several datasets show that the method achieves better results than the state-of-the-art, it can be computed faster, and it provides an intuitive interpretation for the deformation modes.

Sami Sebastian Brandt, Hanno Ackermann, Stella Grasshof

We propose a general, prior-free approach for the uncalibrated non-rigid structure-from-motion problem for modelling and analysis of non-rigid objects such as human faces. The word general refers to an approach that recovers the non-rigid affine structure and motion from 2D point correspondences by assuming that (1) the non-rigid shapes are generated by a linear combination of rigid 3D basis shapes, (2) that the non-rigid shapes are affine in nature, i.e., they can be modelled as deviations from the mean, rigid shape, (3) and that the basis shapes are statistically independent. In contrast to the majority of existing works, no prior information is assumed for the structure and motion apart from the assumption the that underlying basis shapes are statistically independent. The independent 3D shape bases are recovered by independent subspace analysis (ISA). Likewise, in contrast to the most previous approaches, no calibration information is assumed for affine cameras; the reconstruction is solved up to a global affine ambiguity that makes our approach simple but efficient. In the experiments, we evaluated the method with several standard data sets including a real face expression data set of 7200 faces with 2D point correspondences and unknown 3D structure and motion for which we obtained promising results.

Christoph Reinders, Hanno Ackermann, Michael Ying Yang et al.

In recent years, data-driven methods have shown great success for extracting information about the infrastructure in urban areas. These algorithms are usually trained on large datasets consisting of thousands or millions of labeled training examples. While large datasets have been published regarding cars, for cyclists very few labeled data is available although appearance, point of view, and positioning of even relevant objects differ. Unfortunately, labeling data is costly and requires a huge amount of work. In this paper, we thus address the problem of learning with very few labels. The aim is to recognize particular traffic signs in crowdsourced data to collect information which is of interest to cyclists. We propose a system for object recognition that is trained with only 15 examples per class on average. To achieve this, we combine the advantages of convolutional neural networks and random forests to learn a patch-wise classifier. In the next step, we map the random forest to a neural network and transform the classifier to a fully convolutional network. Thereby, the processing of full images is significantly accelerated and bounding boxes can be predicted. Finally, we integrate data of the Global Positioning System (GPS) to localize the predictions on the map. In comparison to Faster R-CNN and other networks for object recognition or algorithms for transfer learning, we considerably reduce the required amount of labeled data. We demonstrate good performance on the recognition of traffic signs for cyclists as well as their localization in maps.

Florian Kluger, Hanno Ackermann, Michael Ying Yang et al.

We present a novel approach for vanishing point detection from uncalibrated monocular images. In contrast to state-of-the-art, we make no a priori assumptions about the observed scene. Our method is based on a convolutional neural network (CNN) which does not use natural images, but a Gaussian sphere representation arising from an inverse gnomonic projection of lines detected in an image. This allows us to rely on synthetic data for training, eliminating the need for labelled images. Our method achieves competitive performance on three horizon estimation benchmark datasets. We further highlight some additional use cases for which our vanishing point detection algorithm can be used.

Bastian Wandt, Hanno Ackermann, Bodo Rosenhahn

This paper deals with motion capture of kinematic chains (e.g. human skeletons) from monocular image sequences taken by uncalibrated cameras. We present a method based on projecting an observation into a kinematic chain space (KCS). An optimization of the nuclear norm is proposed that implicitly enforces structural properties of the kinematic chain. Unlike other approaches our method does not require specific camera or object motion and is not relying on training data or previously determined constraints such as particular body lengths. The proposed algorithm is able to reconstruct scenes with limited camera motion and previously unseen motions. It is not only applicable to human skeletons but also to other kinematic chains for instance animals or industrial robots. We achieve state-of-the-art results on different benchmark data bases and real world scenes.

Stefan Siersdorfer, Philipp Kemkes, Hanno Ackermann et al.

Social network analysis is leveraged in a variety of applications such as identifying influential entities, detecting communities with special interests, and determining the flow of information and innovations. However, existing approaches for extracting social networks from unstructured Web content do not scale well and are only feasible for small graphs. In this paper, we introduce novel methodologies for query-based search engine mining, enabling efficient extraction of social networks from large amounts of Web data. To this end, we use patterns in phrase queries for retrieving entity connections, and employ a bootstrapping approach for iteratively expanding the pattern set. Our experimental evaluation in different domains demonstrates that our algorithms provide high quality results and allow for scalable and efficient construction of social graphs.

Michael Ying Yang, Hanno Ackermann, Weiyao Lin et al.

In this paper, we propose a new framework for segmenting feature-based moving objects under affine subspace model. Since the feature trajectories in practice are high-dimensional and contain a lot of noise, we firstly apply the sparse PCA to represent the original trajectories with a low-dimensional global subspace, which consists of the orthogonal sparse principal vectors. Subsequently, the local subspace separation will be achieved via automatically searching the sparse representation of the nearest neighbors for each projected data. In order to refine the local subspace estimation result and deal with the missing data problem, we propose an error estimation to encourage the projected data that span a same local subspace to be clustered together. In the end, the segmentation of different motions is achieved through the spectral clustering on an affinity matrix, which is constructed with both the error estimation and sparse neighbors optimization. We test our method extensively and compare it with state-of-the-art methods on the Hopkins 155 dataset and Freiburg-Berkeley Motion Segmentation dataset. The results show that our method is comparable with the other motion segmentation methods, and in many cases exceed them in terms of precision and computation time.

Michael Ying Yang, Wentong Liao, Hanno Ackermann et al.

Scene understanding is a popular and challenging topic in both computer vision and photogrammetry. Scene graph provides rich information for such scene understanding. This paper presents a novel approach to infer such relations and then to construct the scene graph. Support relations are estimated by considering important, previously ignored information: the physical stability and the prior support knowledge between object classes. In contrast to previous methods for extracting support relations, the proposed approach generates more accurate results, and does not require a pixel-wise semantic labeling of the scene. The semantic scene graph which describes all the contextual relations within the scene is constructed using this information. To evaluate the accuracy of these graphs, multiple different measures are formulated. The proposed algorithms are evaluated using the NYUv2 database. The results demonstrate that the inferred support relations are more precise than state-of-the-art. The scene graphs are compared against ground truth graphs.

Hanno Ackermann, Michael Ying Yang, Bodo Rosenhahn

Sparse subspace clustering (SSC) is an elegant approach for unsupervised segmentation if the data points of each cluster are located in linear subspaces. This model applies, for instance, in motion segmentation if some restrictions on the camera model hold. SSC requires that problems based on the $l_1$-norm are solved to infer which points belong to the same subspace. If these unknown subspaces are well-separated this algorithm is guaranteed to succeed. The algorithm rests upon the assumption that points on the same subspace are well spread. The question what happens if this condition is violated has not yet been investigated. In this work, the effect of particular distributions on the same subspace will be analyzed. It will be shown that SSC fails to infer correct labels if points on the same subspace fall into more than one cluster.