Refine
Year of publication
Document Type
- Article (16)
- Conference Proceeding (9)
- Preprint (1)
Language
- English (26)
Keywords
- Computer Networks and Communications (5)
- Hardware and Architecture (5)
- Software (4)
- General Computer Science (3)
- Media Technology (3)
- Computer Graphics and Computer-Aided Design (2)
- Computer Vision and Pattern Recognition (2)
- Electrical and Electronic Engineering (2)
- Applied Mathematics (1)
- Artificial Intelligence (1)
Institute
- Fakultät für Angewandte Informatik (26)
- Institut für Informatik (26)
- Lehrstuhl für Menschzentrierte Künstliche Intelligenz (25)
- Nachhaltigkeitsziele (3)
- Ziel 3 - Gesundheit und Wohlergehen (3)
- Institut für Zivilrecht (1)
- Juristische Fakultät (1)
- Lehrstuhl für Bürgerliches Recht, Haftungsrecht und Recht der Digitalisierung (1)
- Lehrstuhl für Maschinelles Lernen und Maschinelles Sehen (1)
In this paper, we present a new low-cost system for surface electromyogram (sEMG) acquisition. developed and designed for rehabilitation application purposes. The noninvasive device delivers four-channel EMG bio-signals describing the electrical activity for the right upper limb muscles. The recorded EMG signals obtained from several healthy subjects were exploited to build a database for movement detection and to evaluate the mechanical properties of the upper limb muscles. The proposed study focuses mainly on the influence of the use of the principal component analysis (PCA) method on the movement classification performance based on the sEMG extracted signals. Several tests were conducted, and the simulation results clearly showed the positive impact of PCA as a dimensionality reduction approach with respect to two performance metrics: the classification rate (CR) and the system’s response time. This advantage was confirmed via numerical tests using three different classifiers: K-nearest neighbor (KNN), probabilistic neural network (PNN), and learning vector quantization (LVQ), with and without PCA. The obtained classification rates highlighted the success of the proposed method since a clear improvement in the classification rates was achieved.
A correct food tray sealing is required to preserve food properties and safety for consumers. Traditional food packaging inspections are made by human operators to detect seal defects. Recent advances in the field of food inspection have been related to the use of hyperspectral imaging technology and automated vision-based inspection systems. A deep learning-based approach for food tray sealing fault detection using hyperspectral images is described. Several pixel-based image fusion methods are proposed to obtain 2D images from the 3D hyperspectral image datacube, which feeds the deep learning (DL) algorithms. Instead of considering all spectral bands in region of interest around a contaminated or faulty seal area, only relevant bands are selected using data fusion. These techniques greatly improve the computation time while maintaining a high classification ratio, showing that the fused image contains enough information for checking a food tray sealing state (faulty or normal), avoiding feeding a large image datacube to the DL algorithms. Additionally, the proposed DL algorithms do not require any prior handcraft approach, i.e., no manual tuning of the parameters in the algorithms are required since the training process adjusts the algorithm. The experimental results, validated using an industrial dataset for food trays, along with different deep learning methods, demonstrate the effectiveness of the proposed approach. In the studied dataset, an accuracy of 88.7%, 88.3%, 89.3%, and 90.1% was achieved for Deep Belief Network (DBN), Extreme Learning Machine (ELM), Stacked Auto Encoder (SAE), and Convolutional Neural Network (CNN), respectively.
This paper explores face anonymization techniques in the context of the General Data Protection Regulation (GDPR) amidst growing privacy concerns due to the widespread use of personal data in machine learning. We focus on unstructured data, specifically facial data, and discuss two approaches to assessing re-identification risks: the risk- based approach supported by GDPR and the zero or strict approach. Emphasizing a process-oriented perspective, we argue that face anonymization should consider the overall data processing context, including the actors involved and the measures taken, to achieve legally secure anonymization under GDPR’s stringent requirements.
When pharmaceutical treatments for Parkinson’s Disease (PD) are no longer effective, Deep Brain Stimulation (DBS) surgery, a procedure that entails the stimulation of the Subthalamic Nucleus (STN), is another treatment option. However, the success rate of this surgery heavily relies on the precise location of the STN, as well as the correct positioning of the stimulation electrode. In order to ensure the correct location, Micro-Electrode Recordings (MERs) are analyzed. During surgery, MERs capture brain signals while inserted in the brain, receiving different brain activity depending on the crossed brain area. The location of the STN is guaranteed when brain signals from MERs meet certain criteria. Nevertheless, MER signals are sensitive to various artifacts coming from machinery or other electrical equipment in the operating theater; patient activity; and electrode motion. These all lower the signal-to-noise ratio of the MER signals. MER signals are stochastic, multicomponent, transient, and non-stationary in nature, and they contain multi-unit neural activity in the form of spikes and artefacts. Thus, accurately defining that MERs are located in the STN is not an easy task. This work analyzes relevant features from MER, based on analyzing spike activity and local field signals. Six different classification algorithms are used, together with the optimal input feature selection. The algorithms are trained using supervised Leave-One-Out Cross-Validation. MER data were collected in a real scenario from 14 PD patients during DBS implantation surgery. The dataset is publicly available. The results derived from the use of this method show an accuracy of up to 100% in detecting where the MER electrode is located in the STN brain area.
Breast cancer is a prevalent disease that primarily affects women globally, but it can also affect men. Early detection is crucial for better treatment outcomes and mammography is a common screening method. Recommendations for mammograms vary by age and country. Early breast-cancer screening is vital for timely interventions. This paper aims to introduce artificial-intelligence methods through deep-learning approaches utilizing pre-trained CNN-based models for the diagnosis of masses depicted in breast images. These masses may be either malignant or benign, necessitating distinct management strategies for each scenario. The experiments conducted on pre-trained models (AlexNet, InceptionV3 and ResNet18) are designed to underscore the significance of selecting the batch size and adaptive learning rate in influencing the results, ultimately facilitating a notable enhancement in classification rates. Pre-trained models applied to a merged dataset comprising three datasets (Inbreast+MIAS+DDSM) yielded an accuracy of 93.7% for InceptionV3 and 88.9% for AlexNet. However, the most favorable outcome was observed with ResNet18, achieving an accuracy of 95% (with precision, recall and F1-score of 94.90%, 94.91% and 94.91%, respectively).
GANonymization: a GAN-based face anonymization framework for preserving emotional expressions
(2025)
In recent years, the increasing availability of personal data has raised concerns regarding privacy and security. One of the critical processes to address these concerns is data anonymization, which aims to protect individual privacy and prevent the release of sensitive information. This research focuses on the importance of face anonymization. Therefore, we introduce GANonymization, a novel face anonymization framework with facial expression-preserving abilities. Our approach is based on a high-level representation of a face, which is synthesized into an anonymized version based on a generative adversarial network (GAN). The effectiveness of the approach was assessed by evaluating its performance in removing identifiable facial attributes to increase the anonymity of the given individual face. Additionally, the performance of preserving facial expressions was evaluated on several affect recognition datasets and outperformed the state-of-the-art methods in most categories. Finally, our approach was analyzed for its ability to remove various facial traits, such as jewelry, hair color, and multiple others. Here, it demonstrated reliable performance in removing these attributes. Our results suggest that GANonymization is a promising approach for anonymizing faces while preserving facial expressions.
