A5026954083- Advanced Neural Network Applications
- Adversarial Robustness in Machine Learning
- Autonomous Vehicle Technology and Safety
- Vehicle License Plate Recognition
- Anomaly Detection Techniques and Applications
- Integrated Circuits and Semiconductor Failure Analysis
Zhengzhou University of Light Industry
2023-2024
Deep neural networks are extremely vulnerable to attacks and threats from adversarial examples. These examples deliberately crafted by attackers can easily fool classification models adding imperceptibly tiny perturbations on clean images. This brings a great challenge image security for deep learning. Therefore, studying designing attack algorithms generating is essential building robust models. Moreover, transferable in that they mislead multiple different classifiers across makes...
In black-box scenarios, most transfer-based attacks usually improve the transferability of adversarial examples by optimizing gradient calculation input image. Unfortunately, since information is only calculated and optimized for each pixel point in image individually, generated tend to overfit local model have poor target model. To tackle issue, we propose a resize-invariant method (RIM) logical ensemble transformation (LETM) enhance examples. Specifically, RIM inspired property Deep Neural...
Lane mark detection is an important task for autonomous driving. Many researchers have proposed many models. But the driving environment much more complex, especially some challenging scenarios, such as vehicle occlusion, severe degradation, heavy shadow, and so on. It difficult to detect lane in a limited local receptive field under above scenarios. For that reason, we propose network based on multihead self-attention. can find spatial relationships among points global viewpoint enlarge its...
Lane detection is an important component of advanced driving aided system (ADAS). It a combined the planning and control algorithms. Therefore, it has high standards for accuracy speed. Recently several researchers have worked extensively on this topic. An increasing number been interested in self-attention-based lane detection. In difficult situations such as shadows, bright lights, nights extracting global information effective. Regardless channel or spatial attention, cannot independently...