arXiv subjects

David Brunner

Publications and source records attributed to David Brunner.

7 recordsLinked to original sources

arXiv2026

BMCTrack-d: Pig re-identification and tracking via back marks in challenging camera settings

Automated pig monitoring is essential for assessing their health, behaviour, and welfare. To date, most pig monitoring solutions operate on the group-level, because individual-level monitoring requires reliable long-term identification and tracking of each animal. For domesticated pigs this remains challenging because pigs of the same breed often have highly uniform appearances. Moreover, research on pig monitoring is almost exclusively reported in top-down view camera settings, which considerably ease tracking, but are not always an option in practice. In this work, BMCTrack-d is presented, a novel tracking-by-detection approach that leverages unique back marks to enable robust pig re-identification and tracking in a challenging side-view camera setting, afflicted by rapidly moving pigs, severe occlusions and low resolution. The method first predicts the detected pigs' identities using a neural network-based back mark classifier. To improve re-identification reliability over time, two dedicated post-processing stages are introduced: a temporal prediction consistency check, which validates the identity assignments against the recent prediction history, and deduplication, which resolves conflicting identity assignments in each time step. By explicitly prioritising accurate, appearance-based re-identification over continuous tracking, the proposed approach addresses a key limitation of existing trackers for individual-level monitoring scenarios. On a demanding test set BMCTrack-d outperforms two strong baselines, BoT-SORT-ReID and TrackTrack-ReID, by 9.11% and 1.03%, respectively, in higher-order tracking accuracy. These results demonstrate the effectiveness of back mark-based re-identification and tracking for robust individual-level pig monitoring in challenging settings.

cs.CV↗

arXiv2026

Insights on back marking for the automated identification of animals

To date, there is little research on how to design back marks to best support individual-level monitoring of uniform looking species like pigs. With the recent surge of machine learning-based monitoring solutions, there is a particular need for guidelines on the design of marks that can be effectively recognised by such algorithms. This study provides valuable insights on effective back mark design, based on the analysis of a machine learning model, trained to distinguish pigs via their back marks. Specifically, a neural network of type ResNet-50 was trained to classify ten pigs with unique back marks. The analysis of the model's predictions highlights the significance of certain design choices, even in controlled settings. Most importantly, the set of back marks must be designed such that each mark remains unambiguous under conditions of motion blur, diverse view angles and occlusions, caused by animal behaviour. Further, the back mark design must consider data augmentation strategies commonly employed during model training, like colour, flip and crop augmentations. The generated insights can support individual-level monitoring in future studies and real-world applications by optimizing back mark design.

cs.CV↗

arXiv2024

Data Distribution Shifts in (Industrial) Federated Learning as a Privacy Issue

We consider industrial federated learning, a collaboration between a small number of powerful, potentially competing industrial players, mediated by a third party aspiring to improve the service it provides to its customers. We argue that this configuration harbours covert privacy risks that do not arise in e.g. cross-device settings. Companies are very protective of their intellectual property and production processes. Information about changes to their production and the timing of which is to be kept private. We study a scenario in which one of the collaborators infers changes to their competitors' production by detecting potentially subtle temporal data distribution shifts. In this framing, a data distribution shift is always problematic, even if it has no negative effect on training convergence. Thus, our goal is to find means that allow the detection of distributional shifts better than customary evaluation metrics. Based on the assumption that even minor shifts translate into the collaboratively learned machine learning model, the attacker tracks the shared models' internal state with a selection of metrics from literature in order to pick up on relevant changes. In an empirical study on benchmark datasets, we show an honest-but-curious attacker to be capable of detecting subtle distributional shifts on other clients, in some cases long before they become obvious in evaluation.

cs.LG↗

arXiv2024

Pilot-Based SFO Estimation for Bistatic Integrated Sensing and Communication

Enabling bistatic radar sensing within the context of integrated sensing and communication (ISAC) for future sixth generation mobile networks demands strict synchronization accuracy, which is particularly challenging to be achieved with over-the-air synchronization. Existing algorithms handle time and frequency offsets adequately, but provide insufficiently accurate sampling frequency offset (SFO) estimates that result in degradation of obtained radar images in the form of signal-to-noise ratio loss and migration of range and Doppler shift. This article introduces an SFO estimation algorithm named tilt inference of time offset (TITO) for orthogonal frequency-division multiplexing (OFDM)-based ISAC. Using available pilot subcarriers, TITO obtains channel impulse response estimates and extracts information on the SFO-induced delay migration to a dominant reference path with constant range, Doppler shift, and angle between transmit and receive ISAC nodes. TITO then adaptively selects the delay estimates that are only negligibly impaired by SFO-induced intersymbol interference, ultimately employing them to estimate the SFO. Assuming a scenario without a direct line-of-sight (LoS) between the aforementioned transmitting and receiving ISAC nodes, a system concept with a relay reflective intelligent surface (RIS) is used to create the aforementioned reference path is proposed. Besides a mathematical derivation of accuracy bounds, simulation and measurements at 26.2 GHz are presented to demonstrate TITO's superiority over existing methods in terms of SFO estimation accuracy and robustness.

eess.SP↗

arXiv2024

Bistatic OFDM-based ISAC with Over-the-Air Synchronization: System Concept and Performance Analysis

Integrated sensing and communication (ISAC) has been defined as one goal for 6G mobile communication systems. In this context, this article introduces a bistatic ISAC system based on orthogonal frequency-division multiplexing (OFDM). While the bistatic architecture brings advantages such as not demanding full duplex operation with respect to the monostatic one, the need for synchronizing transmitter and receiver is imposed. In this context, this article introuces a bistatic ISAC signal processing framework where an incoming OFDM-based ISAC signal undergoes over-the-air synchronization based on preamble symbols and pilots. Afterwards, bistatic radar processing is performed using either only pilot subcarriers or the full OFDM frame. The latter approach requires estimation of the originally transmitted frame based on communication processing and therefore error-free communication, which can be achieved via appropriate channel coding. The performance and limitations of the introduced system based on both aforementioned approaches are assessed via an analysis of the impact of residual synchronization mismatches and data decoding failures on both communication and radar performances. Finally, the performed analyses are validated by proof-of-concept measurement results.

eess.SP↗

arXiv2023

Bistatic OFDM-based Joint Radar-Communication: Synchronization, Data Communication and Sensing

This article introduces a bistatic joint radar-communication (RadCom) system based on orthogonal frequency-division multiplexing (OFDM). In this context, the adopted OFDM frame structure is described and system model encompassing time, frequency, and sampling synchronization mismatches between the transmitter and receiver of the bistatic system is outlined. Next, the signal processing approaches for synchronization and communication are discussed, and radar sensing processing approaches using either only pilots or a reconstructed OFDM frame based on the estimated receive communication data are presented. Finally, proof-of-concept measurement results are presented to validate the investigated system and a trade-off between frame size and the performance of the aforementioned processing steps is observed.

eess.SP↗

arXiv2020

Experimental Evaluation of Vibration Influence on a Resonant MEMS Scanning System for Automotive Lidars

This paper demonstrates a vibration test for a resonant MEMS scanning system in operation to evaluate the vibration immunity for automotive lidar applications. The MEMS mirror has a reinforcement structure on the backside of the mirror, causing vibration coupling by a mismatch between the center of mass and the rotation axis. An analysis of energy variation is proposed, showing direction dependency of vibration coupling. Vibration influences are evaluated by transient vibration response and vibration frequency sweep using a single tone vibration for translational y- and z- axis. The measurement results demonstrate standard deviation (STD) amplitude and frequency errors are up to 1.64 % and 0.26 %, respectively, for 2 grms single tone vibrations on y axis. The simulation results also show a good agreement with both measurements, proving the proposed vibration coupling mechanism of the MEMS mirror. The phased locked loop (PLL) improves the STD amplitude and frequency errors to 0.91 % and 0.15 % for y axis vibration, corresponding to 44.4 % and 43.0 % reduction, respectively, showing the benefit of a controlled MEMS mirror for reliable automotive MEMS lidars.

eess.SY↗