Stroke-induced hand motor dysfunction severely limits activities of daily living (ADL). While conventional systems face challenges in portability and sustained actuation accuracy, this work addresses these limitations through an integrated adaptive control framework and a lightweight 10-degrees-of-freedom (DoFs) tendon-driven exoskeleton. The system employs a rigid–flexible coupling design with a wearable mass under 300 g, ensuring compatibility across various finger lengths. The system is implemented via a motor imagery-based brain–computer interface (MI-BCI); by processing 64-channel electroencephalogram (EEG) signals, the system adaptively maps motor intent into three discrete grasp intensity levels (20%, 50%, and 80% maximum voluntary contraction). To reduce cognitive load and enhance system stability during rehabilitation, we propose a novel “Force–Topology Coupling” control paradigm. This paradigm functions as a synergistic filter, leveraging the correlation between intended effort level (IEL) and grasp taxonomy to map intensity levels to ADL-specific grasps (lateral, precision, and power). Validation with healthy subjects demonstrated 0° to 90° joint mobility and the successful execution of 9 ADL tasks. The results verify the efficacy of utilizing adaptive MI-BCI modulation to trigger biomechanically precise assistance, establishing a foundational computational paradigm with significant potential for clinical stroke rehabilitation.

En los últimos años los algoritmos de inteligencia artificial han demostrado tener grandes resultados en diferentes áreas de aplicación, tales como robótica, medicina, seguridad informática, finanzas, entre otras. En el contexto de procesamiento digital de imágenes, el uso del aprendizaje profundo está siendo aplicado para remover marcas de agua visibles en imágenes digitales, con la finalidad de eliminar la protección de derechos de autor de los propietarios de las imágenes en cuestión. El presente trabajo realiza una recopilación de los trabajos más recientes que remueven marcas de agua visibles a través de aprendizaje profundo, con la finalidad de analizar las tendencias actuales que permitan diseñar algoritmos de marcado de agua visible más robustos ante este tipo de herramientas de remoción.

To meet daily grasping needs under lightweight, low-complexity wearable constraints, this study proposes an underactuated multi-finger prosthetic hand with transmission–control co-design to achieve predictable multi-joint synergies and stable grasps under limited actuation. The prototype uses six miniature motors to drive 14 joint degrees of freedom (DOFs): four fingers have active metacarpophalangeal actuation with tendon-driven underactuated proximal and distal interphalangeal joints, while the thumb provides two independently controlled DOFs for opposition expansion and posture adjustment. It supports five-finger power grasps, tripod pinches, and lateral pinches. To mitigate tendon slack and stroke inconsistency, active/passive tendon-length constraints are defined, and an equal-stroke configuration is obtained via chord-to-arc mapping. A layered STM32F767-based controller combines a reference rotation range limit (free motion) with encoder speed-decay detection (contact/near-stall) to realize per-finger termination and overdrive protection without force/tactile sensors. Experiments report a total mass of 176.6 g and a peak single-finger driving force of approximately 2.8 N. Following the Feix GRASP taxonomy (33 types), the hand reproduces 24 types (72.7%), covering power, intermediate and precision grasps, both thumb abduction/adduction postures, and palm–pad–side opposition/contact, with stable grasp formation across objects of varying geometries.

In recent years, several proposals have emerged for executing arithmetic operations using different variants of Spiking Neural P (SNP) systems. However, some of these proposals rely on distinct circuits for each arithmetic operation, while others mandate preliminary configurations for result computation. Recent research suggests that the biological brain decides to activate or inhibit specific neurons based on the operations performed, without prior preparation. Building upon this understanding, the current work introduces a real-number arithmetic SNP circuit capable of dynamically adjusting its behavior without the need for prior configuration. This adaptability is achieved by selecting between addition or subtraction through the utilization of astrocyte-like control and colored spikes. To validate its performance, the circuit was implemented on an FPGA system. The results indicate that the growth in the quantity of 10th-order digits is comparable to recent proposals in terms of hardware usage, requiring fewer neurons than alternative approaches. Moreover, the computation of floating-point numbers enhances the resolution and precision in various arithmetic applications.