No abstract

In this paper we propose a method based on deep learning that detects multiple people from a single overhead depth image with high reliability. Our neural network, called DPDnet, is based on two fully-convolutional encoder-decoder neural blocks based on residual layers. The main block takes a depth image as input and generates a pixel-wise confidence map, where each detected person in the image is represented by a Gaussian-like distribution. The refinement block combines the depth image and the output from the main block, to refine the confidence map. Both blocks are simultaneously trained end-to-end using depth images and head position labels.The experimental work shows that DPDnet outperforms state-of-the-art methods, with accuracies greater than 99% in three different publicly available datasets, without retraining not fine-tuning. In addition, the computational complexity of our proposal is independent of the number of people in the scene and runs in real time using conventional GPUs.

PhotoPlethysmoGraphic (PPG) signal is an easily accessible biological signal that gives valuable diagnostic information. The novelty is the study procedure of the dynamic of the PPG signals, in our case of young and healthy individuals, with Deep Neural Network, which allows finding the dynamic behavior at different timescales. On a small timescale, the dynamic behavior of the PPG signal is predominantly quasi-periodic. On a large timescale, a more complex dynamic diversity emerges, but never a chaotic behavior as earlier studies had reported. The procedure that determines the dynamics of the PPG signal consists of contrasting the dynamics of a PPG signal with well-known dynamics-named reference signals in this study-, mostly present in physical systems, such as periodic, quasi-periodic, aperiodic, chaotic or random dynamics. For this purpose, this paper provides two methods of analysis based on Deep Neural Network (DNN) architectures. The former uses a Convolutional Neural Network (CNN) architecture model. Upon training with reference signals, the CNN model identifies the dynamics present in the PPG signal at different timescales, assigning, according to a classification process, an occurrence probability to each of them. The latter uses a Recurrent Neural Network (RNN) based on a Long Short-Term Memory (LSTM) architecture. With each of the signals, whether reference signals or PPG signals, the RNN model infers an evolution function (nonlinear regression model) based on training data, and considers its predictive capability over a relatively short time horizon. INDEX TERMS Biological signal, DNN architectures, PPG signal dynamic, timescales.

No abstract

Shape-from-Template (SfT) solves the registration and 3D reconstruction of a deformable 3D object, represented by the template, from a single image. Recently, methods based on deep learning have been able to solve SfT for the wide-baseline case in real-time, clearly surpassing classical methods. However, the main limitation of current methods is the need for fine tuning of the neural models to a specific geometry and appearance represented by the template texture map. We propose the first texture-generic deep learning SfT method which adapts to new texture maps at run-time, without the need for texture specific fine tuning. We achieve this by dividing the problem into a segmentation step and a registration and reconstruction step, both solved with deep learning. We include the template texture map as one of the neural inputs in both steps, training our models to adapt to different ones. We show that our method obtains comparable or better results to previous deep learning models, which are texture specific. It works in challenging imaging conditions, including complex deformations, occlusions, motion blur and poor textures. Our implementation runs in real-time, with a low-cost GPU and CPU.