Coronary artery disease is the most frequent type of heart disease caused by an abnormal narrowing of coronary arteries, also called stenosis or atherosclerosis. It is also the leading cause of death globally. Currently, X-ray Coronary Angiography (XCA) remains the gold-standard imaging technique for medical diagnosis of stenosis and other related conditions. This paper presents a new method for the automatic detection of coronary artery stenosis in XCA images, employing a pre-trained (VGG16, ResNet50, and Inception-v3) Convolutional Neural Network (CNN) via Transfer Learning. The method is based on a network-cut and fine-tuning approach. The optimal cut and fine-tuned layers were selected following 20 different configurations for each network. The three networks were fine-tuned using three strategies: only real data, only artificial data, and artificial with real data. The synthetic dataset consists of 10,000 images ($80\%$ for training, $20\%$ for validation) produced by a generative model. These different configurations were analyzed and compared using a real dataset of 250 real XCA images (125 for testing and 125 for fine-tuning), regarding their randomly initiated CNNs and a fourth custom CNN, trained as well with artificial and real data. The results showed that pre-trained VGG16, ResNet50, and Inception-v3 cut on an early layer and fine-tuned, overcame the referencing CNNs performance. Specifically, Inception-v3 provided the best stenosis detection with an accuracy of $0.95$, a precision of $0.93$, sensitivity, specificity, and $F_1$ score of $0.98$, $0.92$, and $0.95$, respectively. Moreover, a class activation map is applied to identify the high attention regions for stenosis detection.

Advances in the field of Brain-Computer Interfaces (BCIs) aim, among other applications, to improve the movement capacities of people suffering from the loss of motor skills. The main challenge in this area is to achieve real-time and accurate bio-signal processing for pattern recognition, especially in Motor Imagery (MI). The significant interaction between brain signals and controllable machines requires instantaneous brain data decoding. In this study, an embedded BCI system based on fist MI signals is developed. It uses an Emotiv EPOC+ Brainwear^®, an Altera SoCKit^® development board, and a hexapod robot for testing locomotion imagery commands. The system is tested to detect the imagined movements of closing and opening the left and right hand to control the robot locomotion. Electroencephalogram (EEG) signals associated with the motion tasks are sensed on the human sensorimotor cortex. Next, the SoCKit processes the data to identify the commands allowing the controlled robot locomotion. The classification of MI-EEG signals from the F3, F4, FC5, and FC6 sensors is performed using a hybrid architecture of Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks. This method takes advantage of the deep learning recognition model to develop a real-time embedded BCI system, where signal processing must be seamless and precise. The proposed method is evaluated using k-fold cross-validation on both created and public Scientific-Data datasets. Our dataset is comprised of 2400 trials obtained from four test subjects, lasting three seconds of closing and opening fist movement imagination. The recognition tasks reach 84.69% and 79.2% accuracy using our data and a state-of-the-art dataset, respectively. Numerical results support that the motor imagery EEG signals can be successfully applied in BCI systems to control mobile robots and related applications such as intelligent vehicles.