We introduce a novel, fabric-based, flexible, and stretchable tactile sensor, which is capable of seamlessly covering natural shapes. As humans robots have curved body parts that move with respect to each other, the practical usage traditional rigid sensor arrays limited. Rather, flexible skin required. Our design allows for several cells be embedded in single patch. It can an arbitrary perimeter cover free-form surfaces. In this article we discuss construction evaluate its performance....

Our goal with this survey is to provide an overview of the state art deep learning methods for face generation and editing using StyleGAN. The covers evolution StyleGAN, from PGGAN StyleGAN3, explores relevant topics such as suitable metrics training, different latent representations, GAN inversion spaces image editing, cross-domain stylization, restoration, even Deepfake applications. We aim entry point into field readers that have basic knowledge about are looking accessible introduction overview.

An extension of T. Kohonen's (1982) self-organizing mapping algorithm together with an error-correction scheme based on the Widrow-Hoff learning rule is applied to develop a for visuomotor coordination simulated robot arm. Learning occurs by sequence trial movements without need external teacher. Using input signals from pair cameras, closed arm system able reduce its positioning error about 0.3% linear dimensions work space. This achieved choosing connectivity three-dimensional lattice...

Orientation-selective cells in the striate cortex of higher animals are organized as a hierarchical topographic map two stimulus features: (i) position visual space and (ii) orientation. We show that observed structure can arise from principle continuous mapping. For realization this we use mathematical model be interpreted an adaptive process changing set synaptic weights, or connection strengths, between layers cells. The patterns orientation preference selectivity generated by similar to...

The P300 component of an event related potential is widely used in conjunction with brain-computer interfaces (BCIs) to translate the subjects intent by mere thoughts into commands control artificial devices. A well known application spelling words while selection letters carried out focusing attention target letter. In this paper, we present a P300-based online BCI which reaches very competitive performance terms information transfer rates. addition, propose method that optimizes rates...

A lot of motion planning research in robotics focuses on efficient means to find trajectories between individual start and goal regions, but it remains challenging specify plan robotic manipulation actions which consist multiple interdependent subtasks. The Task Constructor framework we present this work provides a flexible transparent way define such actions, enhancing the capabilities popular MoveIt!. <sup xmlns:mml="http://www.w3.org/1998/Math/MathML"...

With the availability of more powerful computers it is nowadays possible to perform pixel based operations on real camera images even in full color space. New adaptive classification tools like neural networks make develop special-purpose object detectors that can segment arbitrary objects with a complex distribution feature space after training one or several previously labeled image(s). The paper focuses detailed comparison approach local linear maps (LLMs) classifier normal distributions....

Service robots which directly interact with humans in highly unstructured, unpredictable and dynamic environments must be able to flexibly adapt their motion reaction unforeseen events or obstacles they provide a new feasible trajectory real-time. Hence, algorithms come into focus replan the path its time evolution from arbitrary initial conditions within milliseconds. We present real-time algorithm generate synchronised time-optimal third-order manipulator trajectories complying maximal...

It is shown that for a class of vector quantization processes, related to neural modeling, the asymptotic density Q(x ) levels in one dimension terms input signal distribution P(x) power law Q(x)=C-P(x)(alpha ), where exponent alpha depends on number n neighbors each side unit and given by alpha=2/3-1/(3n (2)+3[n+1](2)). The level calculated, Monte Carlo simulations are presented.

The advent of sensor arrays providing tactile feedback with high spatial and temporal resolution asks for new control strategies to exploit this important valuable sensory channel grasping manipulation tasks.In paper, we introduce a framework realize whole set servoing tasks, i.e. tasks that intend specific interaction pattern.This includes such simple like tracking touched object, maintaining both contact location force, as well more elaborate an object's pose or object...

In this paper, we present a probabilistic spatial approach to build compact 3-D representations of unknown objects probed by tactile sensors. Our exploits the high frame rates provided modern sensors and utilizes Kalman filters model contact point cloud that is efficiently stored in kd-tree. The quality generated shape compared with naive averaging approach, show our method provides superior accuracy. We also evaluate feasibility object classification combining representations, together...

This paper describes GREFIT (Gesture REcognition based on FInger Tips), a neural network-based system which recognizes continuous hand postures from gray-level video images (posture capturing). Our approach yields full identification of all finger joint angles (making, however, some assumptions about couplings to simplify computations). allows reconstruction the three-dimensional (3-D) shape, using an articulated model with 16 segments and 20 angles. uses two-stage solve this task. In first...

Two experiments on the perception and eye-movement scanning of a set six overtly ambiguous pictures are reported. In first experiment it was shown that specific perceptual interpretations an picture usually correlate with parameters gaze-position distributions. second these distributions were used for image processing initial in such way regions which attracted less fixations brightness all elements lowered. The preprocessed then to group 150 naïve subjects identification. results this...

Humans easily recognize where another person is looking and often use this information for interspeaker coordination. We present a method based on three neural networks of the local linear map type which enables computer to identify head orientation user by learning from examples. One network used color segmentation, second localization face, third final recognition orientation. The system works at frame rate one image per common workstation, analyze accuracy achieved different processing...

We propose a new and efficient approach to compute task oriented quality measures for dextrous grasps. Tasks can be specified as single wrench applied, rough direction in form of cone, or complex polytope. Based on the linear matrix inequality formalism treat friction cone constraints we formulate respective convex optimization problems, whose solutions give maximal applicable together with needed contact forces. Numerical experiments show that application grasps many contacts is possible.