Search

Forensic license plate recognition (FLPR) remains an open challenge in legal contexts such as criminal investigations, where unreadable license plates (LPs) need to be deciphered from highly compressed and/or low resolution footage, e.g., from surveillance cameras. In this work, we propose a side-informed Transformer architecture that embeds knowledge on the input compression level to improve recognition under strong compression. We show the effectiveness of Transformers for license plate recognition (LPR) on a low-quality real-world dataset. We also provide a synthetic dataset that includes strongly degraded, illegible LP images and analyze the impact of knowledge embedding on it. The network outperforms existing FLPR methods and standard state-of-the art image recognition models while requiring less parameters. For the severest degraded images, we can improve recognition by up to 8.9 percent points., Accepted at ICIP 2022, Code: https://faui1-gitlab.cs.fau.de/denise.moussa/forensic-license-plate-transformer/

When designing multispectral imaging systems for classifying different spectra it is necessary to choose a small number of filters from a set with several hundred different ones. Tackling this problem by full search leads to a tremendous number of possibilities to check and is NP-hard. In this paper we introduce a novel fast binary search for optimal filter selection that guarantees a minimum distance metric between the different spectra to classify. In our experiments, this procedure reaches the same optimal solution as with full search at much lower complexity. The desired number of filters influences the full search in factorial order while the fast binary search stays constant. Thus, fast binary search allows to find the optimal solution of all combinations in an adequate amount of time and avoids prevailing heuristics. Moreover, our fast binary search algorithm outperforms other filter selection techniques in terms of misclassified spectra in a real-world classification problem.

Employing novel 2D materials with topologically protected current-carrying edge states is promising to boost the on-current in electronic devices. Using nanoribbons is essential to reduce the contribution of the 2D bulk states to the current. Making the nanoribbon widths narrower allows one to put more current-carrying edge states under the gate of a fixed width thus boosting the current. However, the edge states start to interact in narrow nanoribbons. Based on an effective k·p model, we analyze the topologically protected edge states and their conductance for several 2D materials as a function of the normal electric field. We compare the 2D materials MoS 2 , MoSe 2 , WS 2 , and WSe 2 in the topological 1T′ phase and find the largest electric field-induced conductance modulation is in MoS 2 nanoribbons.

Exploiting novel materials with advanced properties is necessary to improve device performance and continue with their scaling for high performance applications at reduced power. Among these materials, topological insulators (TIs) present an exciting opportunity where the highly conductive edge states, which are protected against back scattering, can lead to advances in electronic as well as spin controlled devices. Here, we present first principles results that evaluate topologically protected edge states in MoS 2 nanoribbons. We vary the width of the nanoribbons and show that they transition to trivial insulators below a critical width. Furthermore, the trivial insulator can be a direct or indirect band gap insulator depending on the width and the edge termination. Our results show that including realistic edge terminations can provide valuable insight, especially for narrow width TIs.

In this paper, we propose a data augmentation framework for Optical Character Recognition (OCR). The proposed framework is able to synthesize new viewing angles and illumination scenarios, effectively enriching any available OCR dataset. Its modular structure allows to be modified to match individual user requirements. The framework enables to comfortably scale the enlargement factor of the available dataset. Furthermore, the proposed method is not restricted to single frame OCR but can also be applied to video OCR. We demonstrate the performance of our framework by augmenting a 15% subset of the common Brno Mobile OCR dataset. Our proposed framework is capable of leveraging the performance of OCR applications especially for small datasets. Applying the proposed method, improvements of up to 2.79 percentage points in terms of Character Error Rate (CER), and up to 7.88 percentage points in terms of Word Error Rate (WER) are achieved on the subset. Especially the recognition of challenging text lines can be improved. The CER may be decreased by up to 14.92 percentage points and the WER by up to 18.19 percentage points for this class. Moreover, we are able to achieve smaller error rates when training on the 15% subset augmented with the proposed method than on the original nonaugmented full dataset.

Most state-of-the-art video coders rely on a block structure. For inter-frame prediction, motion vectors are transmitted per block. For example in VVC, the coder can choose between a translational or an affine motion model on a block level, depending on the content. In non-block-based coding, which is on the rise since the development of end-to-end learning based image compression, the motion vectors have to be transmitted differently. Due to the missing inherent block structure, switching between different motion models presents a challenge, but also an opportunity. In this paper, we propose an alternative approach to efficiently signal additional information regarding the motion model to improve the quality of the motion compensated image. Using our methods, we are able to increase the quality of the prediction image in our scenario by 0.40 dB on average and by up to 0.88 dB for sequences with strong and complex motion at the same rate.

We present necessary and sufficient conditions for the interpolation of a set of points and gradient evaluations by a smooth and strongly convex function. The conditions are expressed as quadratic constraints over doubly hyperdom-inant matrices. The current literature focuses on interpolation conditions involving both gradient and function evaluations. An extension of these results to use only gradient evaluations requires factorial growth in the number of constraints. Our approach avoids the factorial growth when the interpolation conditions are applied to both finite step and asymptotic optimization algorithm analysis. We present a numerical example demonstrating that the use of these interpolation conditions in the asymptotic analysis setting can produce tighter bounds than are currently available in the literature.

A method is presented for computing the worst-case disturbance of a given norm for a finite-horizon linear time-varying system with a nonzero initial condition. This method is motivated by the linearized robustness analysis of a nonlinear system about a finite-horizon nominal trajectory. The system is linearized about this trajectory and interconnected with sampled model uncertainties. An additional input term to improve the fidelity of the linear interconnection is introduced and is absorbed into an augmented system with a nonzero initial condition. The method, which analyzes the robustness of the resulting interconnection to disturbance, is demonstrated on the numerical example of a two-link robot arm.