T
Tim Fingscheidt
Researcher at Braunschweig University of Technology
Publications - 250
Citations - 3048
Tim Fingscheidt is an academic researcher from Braunschweig University of Technology. The author has contributed to research in topics: Speech enhancement & Computer science. The author has an hindex of 22, co-authored 231 publications receiving 2140 citations. Previous affiliations of Tim Fingscheidt include Siemens & AT&T Labs.
Papers
More filters
Proceedings ArticleDOI
Beyond the Dcase 2017 Challenge on Rare Sound Event Detection: A Proposal for a More Realistic Training and Test Framework
TL;DR: This paper proposes a rare SED training and test framework, which is reflecting an SED application in a more realistic way, and shows and compares the performance of multi-event (polyphonic) classifiers vs. single-event classifiers while outlining the benefits ofmulti-event training.
Journal ArticleDOI
Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation
TL;DR: This work expands a source-free UDA approach to a continual and therefore online-capable UDA on a single-image basis for semantic segmentation and modifies the source domain statistics in the batch normalization layers, using target domain images in an unsupervised fashion, which yields consistent performance improvements during inference.
Proceedings ArticleDOI
Delayless soft-decision decoding of high-quality audio with adaptively shaped priors
Florian Pflug,Tim Fingscheidt +1 more
TL;DR: A Bayesian framework for delayless full-band soft-decision error concealment for quantized but uncompressed audio utilizing only residual redundancy in the audio signal and channel reliability information is dealt with.
Proceedings ArticleDOI
Towards Tactical Maneuver Detection for Autonomous Driving Based on Vision Only
TL;DR: In this article, the authors used neural networks with temporal convolutions to incorporate temporal information into the detection of tactical maneuvers performed by other traffic participants and achieved an average accuracy of 54.21% with single detection accuracies up to 88.17%.