Intelligent Autonomous Systems (INTAS)

The paper "Structured temporal representation in time series classification with ROCKETs and Hyperdimensional Computing" by Kenny Schlegel, Dmitri A. Rachkovskij, Denis Kleyko,Ross W. Gayler, Peter Protzel, and Peer Neubert has been accepted for publication in the Data Mining and Knowledge Discovery journal.

Abstract: Time series classification poses significant challenges due to the inherent temporal order of the data points and the existence of sequential dependencies between them. The ROCKET family, featuring methods like MiniROCKET, MultiROCKET, and HYDRA, is currently a leading approach in this domain, leveraging convolution kernels to aggregate temporal features into encodings for linear classifiers. However, these models encode temporal features over short temporal windows and then aggregate them as an unordered set of encodings over the longer temporal window of the entire data sequence. This prevents these models from capturing any longer sequence structure. To address this design drawback, we propose integrating hyperdimensional computing into ROCKET methods to explicitly incorporate temporal order of the short-term features within the entire time series. This approach enhances the discriminative power of encodings generated by MiniROCKET, MultiROCKET, and HYDRA where longer-term structure exists in the data, leading to increased classification performance with minimal computational overhead. More specifically, we introduce a method to represent time series as high-dimensional vectors through multiplicative binding of ROCKET encodings with encodings representing temporal order, applying this approach across various ROCKET methods. Additionally, we explore different high-dimensional vector representations of temporal order, yielding diverse similarity kernels that enhance classification accuracy. Through experiments on synthetic datasets, we highlight the limitations of ROCKET methods in handling temporal dependencies and show how the methods based on hyperdimensional computing overcome these limitations. Furthermore, our extensive experimental evaluation with real-world datasets included in the recent UCR archive [1] and additional datasets from [2], validates the advantages of our approach, consistently achieving classification improvements across all ROCKET methods that integrate hyperdimensional computing. Notably, our best model achieves a relative error rate reduction of over 50% compared to the best ROCKET model on several UCR datasets.

The paper "Data-Efficient Spectral Classification of Hyperspectral Data UsingMiniROCKET and HDC-MiniROCKET" by Nick Theisen, Kenny Schlegel, Dietrich Paulus, and Peer Neubert has been accepted at the CASE 2025 conference.

Abstract - The classification of pixel spectra of hyperspectralimages, i. e. spectral classification, is used in many fieldsranging from agricultural, over medical to remote sensing applications and is currently also expanding to areas such asautonomous driving. Even though for full hyperspectral images the best-performing methods exploit spatial-spectral information, performing classification solely on spectral informationhas its own advantages, e. g. smaller model size and thus less data required for training. Moreover, spectral information is complementary to spatial information and improvements on either part can be used to improve spatial-spectral approaches in the future. Recently, 1D-Justo-LiuNet was proposed as a particularly efficient model with very few parameters, which currently defines the state of the art in spectral classification. However, we show that with limited training data the model performance deteriorates. Therefore, we investigate MiniROCKETand HDC-MiniROCKET for spectral classification to mitigatethat problem. The model extracts well-engineered features without trainable parameters in the feature extraction part and is therefore less vulnerable to limited training data. We show that even though MiniROCKET has more parameters it outperforms 1D-Justo-LiuNet in limited data scenarios and ismostly on par with it in the general case.

The paper "On the choice of Vector Symbolic Architectures for Time Series Classification with HDC-MiniROCKET" by Marcel Unger , Kenny Schlegel , Peter Protzel , and Peer Neubert has bee accepted at the IJCNN 2025 conference.

Abstract—Hyperdimensional Computing (HDC) has shown promise in time series classification by enhancing MiniROCKET, forming HDC-MiniROCKET. However, the impact of choosing a specific HDC implementation, referred to as a Vector Symbolic Architecture (VSA), within this framework is so far unexplored. This paper systematically evaluates different VSAs within HDC-MiniROCKET, analyzing their impact on classification performance, hyperparameter sensitivity, and computational efficiency. Our findings reveal that certain VSAs require a significantly broader range of hyperparameter values to achieve optimal accuracy despite similar properties. We demonstrate this effect on a synthetic dataset and validate it across a subset of the real-world benchmark UCR, showing that VSA performance varies depending on the dataset characteristics. Additionally, we investigate the trade-off between computational complexity and classification accuracy. We find that computationally efficient VSAs not only reduce processing time but also achieve comparable or superior accuracy to more complex alternatives. These insights help to select VSAs in HDC-based time series classification.

The paper "Towards Zero-Shot Terrain Traversability Estimation: Challenges and Opportunities" by Ida Germann, Mark O. Mints and Peer Neubert has been accepted at the 1st German Robotics Conference 2025.

Abstract— Terrain traversability estimation is crucial forau tonomous robots, especially in unstructured environments where visual cues and reasoning play a key role. While vision-language models (VLMs) offer potential for zero-shot estimation, the problem of traversability classification remains inherently ill-posed. To explore this, we introduce a small dataset of human-annotated water traversability ratings, revealing that while estimations are subjective, human raters still show some consensus. Additionally, we propose a simple pipeline that integrates VLMs for zero-shot traversability estimation. Our experiments reveal mixed results, suggesting that current foundation models are not yet suitable for practical deployment but provide valuable insights for further research.

After the successful start last year, we will organize another CV-Camp in 2025: https://uni-ko.de/cv-camp

Nick is expert for semantic segmentation and hyperspectal images. He will work in the MineSweeper project.

This new 3-year research project is on visual navigation in challenging and changing environments in the context of civil demining

Congratulations to Jessica Buchner for winning the 1st place jury prize at CV Day for her work "Visual Tracking of Reptiles in a Zoological Context" and to Felix Meyer for winning the 1st place audience award for his work "Camera-based overtaking distance control"!

The Paper "HS3-Bench: A Benchmark and Strong Baseline for Hyperspectral Semantic Segmentation in Driving Scenarios" by Nick Theisen, Robin Bartsch, Dietrich Paulus, and Peer Neubert has been accepted at IROS'24.

Abstract: Semantic segmentation is an essential step for many vision applications in order to understand a scene and the objects within. Recent progress in hyperspectral imaging technology enables the application in driving scenarios and the hope is that the device's perceptive abilities provide an advantage over RGB-cameras. Even though some datasets exist,there is no standard benchmark available to systematically measure progress on this task and evaluate the benefit of hyperspectral data. In this paper, we work towards closing this gap by providing the HyperSpectral Semantic Segmentation benchmark (HS3-Bench). It combines annotated hyperspectral images from three driving scenario datasets and provides standardized metrics, implementations, and evaluation protocols. We use the benchmark to derive two strong baseline models that surpass the previous state-of-the-art performances with and without pre-training on the individual datasets. Further, our results indicate that the existing learning-based methods benefit more from leveraging additional RGB training data than from leveraging the additional hyperspectral channels. This poses important questions for future research on hyperspectral imaging for semantic segmentation in driving scenarios.