Aktuelles

Digital image processing in professional applications places ever-higher demands in terms of compute power and power consumption so that FPGA devices reach their limits. Approximate Computing is an approach to tackle this challenge. It refers to a set of methods to perform calculations not exactly… [weiterlesen]

Two contributions from the joint team of LIS with BMW, PoliTorino, and KIT have been accepted to the DATE 2022 conference. Congratulation to all the members of the team for their joint effort. The first paper entitled "Mind the Scaling Factors: Resilience Analysis of Quantized Adversarially Robust… [weiterlesen]

It is our pleasure to congratulate the winners of the 2021 Xilinx Open Hardware Design Contest in the PYNQ category: Nael Fasfous, Manoj-Rohit Vemparala, and Alexander Frickenstein with their design of "Binary Neural Network-based COVID19 Face-Mask Wear and Positioning Predictor". … [weiterlesen]

Efficiently deploying learning-based systems on embedded hardware is challenging for various reasons, two of which are considered in this paper: The model’s size and its robustness against attacks. Together with our partners from BMW and KIT, we combine adversarial training and model pruning in a… [weiterlesen]

we had got to know Anh Vu Doan in autumn 2017 in Seoul, South Korea at Embedded Systems Week / NOCS Symposium and had very good discussions there, which finally lead to hiring him as a postdoc at LIS from Spring 2018. In the following three years, Anh Vu significantly contributed to a variety of our… [weiterlesen]

In the context of the ongoing COVID-19 pandemic, face masks offer an effective contribution to healthcare. Wearing and positioning the mask correctly is essential for its function. Convolutional neural networks (CNNs) offer an excellent solution for visual face recognition and classification of… [weiterlesen]

In a joint team between BMW and TUM, we investigate the robustness of Convolutional Neural Network (CNN) deployment on embedded systems, particularly the robustness against adversarial attacks in automotive application scenarios. In the paper titled "BreakingBED - Breaking Binary and Efficient Deep… [weiterlesen]