Auflistung nach Schlagwort "Deep Neural Networks"
1 - 3 von 3
Treffer pro Seite
Sortieroptionen
- TextdokumentEvaluation of CNN architectures for gait recognition based on optical flow maps(BIOSIG 2017, 2017) Castro,Francisco M.; Marín-Jiménez,Manuel J.; Guil,Nicolás; López-Tapia,Santiago; de la Blanca,Nicolás PérezThis work targets people identification in video based on the way they walk (i.e.gait) by using deep learning architectures. We explore the use of convolutional neural networks (CNN) for learning high-level descriptors from low-level motion features (i.e.optical flow components). The low number of training samples for each subject and the use of a test set containing subjects different from the training ones makes the search of a good CNN architecture a challenging task.We carry out a thorough experimental evaluation deploying and analyzing four distinct CNN models with different depth but similar complexity. We show that even the simplest CNN models greatly improve the results using shallow classifiers. All our experiments have been carried out on the challenging TUMGAID dataset, which contains people in different covariate scenarios (i.e.clothing, shoes, bags).
- KonferenzbeitragA Novel Approach for Sensor Fusion Object Detection in Waste Sorting: The Case of WEEE(EnviroInfo 2023, 2023) Nazeri, Ali; Plociennik, Christiane; Vogelgesang, Malte; Li, Chanchan; Ruskowski, MartinThis paper investigates the application of AI-based methods for characterizing waste materials in sorting processes. With the increasing use of sensors in waste sorting systems, there is an opportunity to integrate data and improve accuracy. AI methods, such as deep object detection models, have the potential to optimize waste management processes and promote sustainability. This research examines the utilization of Sensor Fusion Object Detection in a multi-sensor sorting system, focusing on two different data fusion methods: concatenation and image mirroring. In the first approach, image data is concatenated with data from a hyperspectral near-infrared camera (NIR) and an inductive sensor, where dimensionality reduction techniques are applied to the data from both sensors. The second approach relies on a specific combination of NIR and inductive sensor data to simulate the format of image data. A Siamese Object Detection architecture is developed to train the model. The real-world testing results show that both approaches improve waste characterization accuracy and reliability by augmenting the models’ mean average precision (mAP). These findings demonstrate the potential for AI-based methods to transform the waste separation and management process, leading to more sustainable practices and resource efficiency.
- KonferenzbeitragSub-byte quantization of Mobile Face Recognition Convolutional Neural Networks(BIOSIG 2022, 2022) Sebastian Bunda, Luuk Spreeuwers and Chris ZeinstraConverting convolutional neural networks such as MobileNets to a full integer representation is already quite a popular method to reduce the size and computational footprint of classification networks but its effect on face recognition networks is relatively unexplored. This work presents a method to reduce the size of MobileFaceNet using sub-byte quantization of the weights and activations. It was found that 8-bit and 4-bit versions of MobileFaceNet can be obtained with 98.68% and 98.63% accuracy on the LFW dataset which reduces the footprint to 25% and 12.5% of the original weights respectively. Using mixed-precision, an accuracy of 98.17% can be achieved whilst requiring only 10% of the original weight footprint. It is expected that with a larger training dataset, higher accuracies can be achieved.