Search

Cameras digitize real-world scenes as pixel intensity values with a limited value range given by the available bits per pixel (bpp). High Dynamic Range (HDR) cameras capture those luminance values in higher resolution through an increase in the number of bpp. Most displays, however, are limited to 8 bpp. Naive HDR compression methods lead to a loss of the rich information contained in those HDR images. In this paper, tone mapping algorithms for thermal infrared images with 16 bpp are investigated that can preserve this information. An optimized multi-scale Retinex algorithm sets the baseline. This algorithm is then approximated with a deep learning approach based on the popular U-Net architecture. The remaining noise in the images after tone mapping is reduced implicitly by utilizing a self-supervised deep learning approach that can be jointly trained with the tone mapping approach in a multi-task learning scheme. Further discussions are provided on denoising and deflickering for thermal infrared video enhancement in the context of tone mapping. Extensive experiments on the public FLIR ADAS Dataset prove the effectiveness of our proposed method in comparison with the state-of-the-art., Comment: 17 pages, 10 figures

Object detection is one of the key tasks in many applications of computer vision. Deep Neural Networks (DNNs) are undoubtedly a well-suited approach for object detection. However, such DNNs need highly adapted hardware together with hardware-specific optimization to guarantee high efficiency during inference. This is especially the case when aiming for efficient object detection in video streaming applications on limited hardware such as edge devices. Comparing vendor-specific hardware and related optimization software pipelines in a fair experimental setup is a challenge. In this paper, we propose a framework that uses a host computer with a host software application together with a light-weight interface based on the Message Queuing Telemetry Transport (MQTT) protocol. Various different target devices with target apps can be connected via MQTT with this host computer. With well-defined and standardized MQTT messages, object detection results can be reported to the host computer, where the results are evaluated without harming or influencing the processing on the device. With this quite generic framework, we can measure the object detection performance, the runtime, and the energy efficiency at the same time. The effectiveness of this framework is demonstrated in multiple experiments that offer deep insights into the optimization of DNNs., Comment: Proceedings of the DAGM German Conference on Pattern Recognition (GCPR) 2022

Multispectral person detection aims at automatically localizing humans in images that consist of multiple spectral bands. Usually, the visual-optical (VIS) and the thermal infrared (IR) spectra are combined to achieve higher robustness for person detection especially in insufficiently illuminated scenes. This paper focuses on analyzing existing detection approaches for their generalization ability. Generalization is a key feature for machine learning based detection algorithms that are supposed to perform well across different datasets. Inspired by recent literature regarding person detection in the VIS spectrum, we perform a cross-validation study to empirically determine the most promising dataset to train a well-generalizing detector. Therefore, we pick one reference Deep Convolutional Neural Network (DCNN) architecture and three different multispectral datasets. The Region Proposal Network (RPN) originally introduced for object detection within the popular Faster R-CNN is chosen as a reference DCNN. The reason is that a stand-alone RPN is able to serve as a competitive detector for two-class problems such as person detection. Furthermore, current state-of-the-art approaches initially apply an RPN followed by individual classifiers. The three considered datasets are the KAIST Multispectral Pedestrian Benchmark including recently published improved annotations for training and testing, the Tokyo Multi-spectral Semantic Segmentation dataset, and the OSU Color-Thermal dataset including recently released annotations. The experimental results show that the KAIST Multispectral Pedestrian Benchmark with its improved annotations provides the best basis to train a DCNN with good generalization ability compared to the other two multispectral datasets. On average, this detection model achieves a log-average Miss Rate (MR) of 29.74 % evaluated on the reasonable test subsets of the three datasets.

In recent years, knapsack problems for (in general non-commutative) groups have attracted attention. In this paper, the knapsack problem for wreath products is studied. It turns out that decidability of knapsack is not preserved under wreath product. On the other hand, the class of knapsack-semilinear groups, where solutions sets of knapsack equations are effectively semilinear, is closed under wreath product. As a consequence, we obtain the decidability of knapsack for free solvable groups. Finally, it is shown that for every non-trivial abelian group $G$, knapsack (as well as the related subset sum problem) for the wreath product $G \wr \mathbb{Z}$ is NP-complete.

In a recent paper we analyzed the space complexity of streaming algorithms whose goal is to decide membership of a sliding window to a fixed language. For the class of regular languages we proved a space trichotomy theorem: for every regular language the optimal space bound is either constant, logarithmic or linear. In this paper we continue this line of research: We present natural characterizations for the constant and logarithmic space classes and establish tight relationships to the concept of language growth. We also analyze the space complexity with respect to automata size and prove almost matching lower and upper bounds. Finally, we consider the decision problem whether a language given by a DFA/NFA admits a sliding window algorithm using logarithmic/constant space., Comment: Extended version of a paper presented at STACS 2018

The computational complexity of the circuit evaluation problem for finite semirings is considered, where semirings are not assumed to have an additive or multiplicative identity. The following dichotomy is shown: If a finite semiring is such that (i) the multiplicative semigroup is solvable and (ii) it does not contain a subsemiring with an additive identity $0$ and a multiplicative identity $1 \neq 0$, then the circuit evaluation problem for the semiring is in $\mathsf{DET} \subseteq \mathsf{NC}^2$. In all other cases, the circuit evaluation problem is $\mathsf{P}$-complete., Comment: Some proof details from the previous version are simplified in the new version

It is shown that the knapsack problem (introduced by Myasnikov, Nikolaev, and Ushakov) is undecidable in a direct product of sufficiently many copies of the discrete Heisenberg group (which is nilpotent of class 2). Moreover, for the discrete Heisenberg group itself, the knapsack problem is decidable. Hence, decidability of the knapsack problem is not preserved under direct products. It is also shown that for every co-context-free group, the knapsack problem is decidable. For the subset sum problem (also introduced by Myasnikov, Nikolaev, and Ushakov) we show that it belongs to the class NL (nondeterministic logspace) for every finitely generated virtually nilpotent group and that there exists a polycyclic group with an NP-complete subset sum problem.

Powerful skew arithmetic circuits are introduced. These are skew arithmetic circuits with variables, where input gates can be labelled with powers $x^n$ for binary encoded numbers $n$. It is shown that polynomial identity testing for powerful skew arithmetic circuits belongs to $\mathsf{coRNC}^2$, which generalizes a corresponding result for (standard) skew circuits. Two applications of this result are presented: (i) Equivalence of higher-dimensional straight-line programs can be tested in $\mathsf{coRNC}^2$; this result is even new in the one-dimensional case, where the straight-line programs produce strings. (ii) The compressed word problem (or circuit evaluation problem) for certain wreath products of finitely generated abelian groups belongs to $\mathsf{coRNC}^2$.

The circuit evaluation problem (also known as the compressed word problem) for finitely generated linear groups is studied. The best upper bound for this problem is $\mathsf{coRP}$, which is shown by a reduction to polynomial identity testing. Conversely, the compressed word problem for the linear group $\mathsf{SL}_3(\mathbb{Z})$ is equivalent to polynomial identity testing. In the paper, it is shown that the compressed word problem for every finitely generated nilpotent group is in $\mathsf{DET} \subseteq \mathsf{NC}^2$. Within the larger class of polycyclic groups we find examples where the compressed word problem is at least as hard as polynomial identity testing for skew arithmetic circuits.

One core challenge of nanoelectromechanical systems (NEMS) is their efficient actuation. A promising concept superseding resonant driving is self-oscillation. Here we demonstrate voltage-sustained self-oscillation of a nanomechanical charge shuttle. Stable transport at 4.2 K is observed for billions of shuttling cycles, giving rise to ohmic current-voltage curves with a sharp dissipation threshold. With only a few nanowatts of input energy the presented scheme is suitable for operation in the millikelvin regime where Coulomb blockade-controlled single electron shuttling is anticipated., Comment: 6 pages, including 4 figures