The minimum torque-change model predicts and reproduces human multi-joint movement data quite well. However, there are three criticisms of the current neural network models for trajectory formation based on the minimu...
详细信息
The minimum torque-change model predicts and reproduces human multi-joint movement data quite well. However, there are three criticisms of the current neural network models for trajectory formation based on the minimum torque-change criteria: (1) their spatial representation of time, (2) back propagation is essential, and (3) they require too many iterations. Accordingly, we propose a new neural network model for trajectory formation based on the minimum torque-change criterion. Our neural network model basically uses a forward dynamics model, an inverse dynamics model, and a trajectory formation mechanism, which generates an approximate minimum torque-change trajectory It does not require spatial representation of time or back propagation. Furthermore, there are less iterations required to obtain an approximate optimal solution. Finally, our neural network model can be broadly applied to the engineering field because it is a new method for solving optimization problems with boundary conditions.
We investigate experimentally the dynamic behaviors of an oscillatory neural network. Computer simulations show an interesting characteristic: the autonomous generation of a limit cycle near a memory pattern (memory r...
详细信息
We investigate experimentally the dynamic behaviors of an oscillatory neural network. Computer simulations show an interesting characteristic: the autonomous generation of a limit cycle near a memory pattern (memory retrieval with ambiguous fluctuation) for an input near it, and of a chaotic orbit wandering among memory patterns (autonomous search) for an input far from them. We also analyze theoretically restricted behaviors near a memory pattern. This type of neural network can treat dynamic informationprocessing of spatiotemporal patterns. As an example, it is shown that continuously transformed pattern cycles for three Japanese characters can be embedded in the limit cycles of the oscillatory neural network by a learning method The characteristic behavior of limit cycles or chaos according to inputs may be useful for developing models of dynamic informationprocessing mechanisms of spatiotemporal patterns.
This paper presents a new approach to the recovery of 3-D structure from multiple pairs of images from different viewpoints. Searching for the corresponding points between images, which is common in stereopsis, is avo...
详细信息
This paper presents a new approach to the recovery of 3-D structure from multiple pairs of images from different viewpoints. Searching for the corresponding points between images, which is common in stereopsis, is avoided. Extracted edges from input images are projected back into 3-D space, and their intersections are calculated directly. Many false intersections may appear, but if we have many pair images, true intersections are extracted by appropriate thresholding. Octree representation of the intersections enables this approach. We consider a way to treat adjacent edge piexels as a line segment rather than as individual points, which differs from previous works and leads to a new algorithm. Experimental results using both synthetic and actual images are also described.
This paper focuses on two areas in our effort to synthesize speech from neuromotor input using neural network models that effect transforms between cognitive intentions to speak, their physiological effects on vocal t...
详细信息
This paper focuses on two areas in our effort to synthesize speech from neuromotor input using neural network models that effect transforms between cognitive intentions to speak, their physiological effects on vocal tract structures, and subsequent realization as acoustic signals. The first area concerns the biomechanical transform between motor commands to muscles and the ensuing articulator behavior. Using physiological data of muscle EMG (electromyography) and articulator movements during natural English speech utterances, three articulator-specific neural networks learn the forward dynamics that relate motor commands to the muscles and motion of the tongue, jaw, and lips. Compared to a fully-connected network, mapping muscle EMG and motion for all three sets of articulators at once, this modular approach has improved performance by reducing network complexity and has eliminated some of the confounding influence of functional coupling among articulators. Network independence has also allowed us to identify and assess the effects of technical and empirical limitations on an articulator-by-articulator basis. This is particularly important for modeling the tongue whose complex structure is very difficult to examine empirically. The second area of progress concerns the transform between articulator motion and the speech acoustics. From the articulatory movement trajectories, a second neural network generates PARCOR (partial correlation) coefficients which are then used to synthesize the speech acoustics. In the current implementation, articulator velocities have been added as the inputs to the network. As a result, the model now follows the fast changes of the coefficients for consonants generated by relatively slow articulatory movements during natural English utterances. Although much work still needs to be done, progress in these areas brings us closer to our goal of emulating speech production processes computationally.
This paper propose a scheme that offers robust extraction of target images in standard from input facial images, in order to realize accurate and automatic identification of human faces. A standard view for target ima...
详细信息
This paper propose a scheme that offers robust extraction of target images in standard from input facial images, in order to realize accurate and automatic identification of human faces. A standard view for target images is defined using internal facial features, i.e., the two eyes and the mouth, as steady reference points of the human face. reliable detection of such facial features is an easy task in practice, the proposed scheme is characterized by combination of two steps: first, all possible regions of facial features are extracted using a color image segmentation algorithm, then the target image is selected from among the candidates defined by tentative combinations of the three reference points, through applying the classification framework using the sub-space method. Preliminary experiments on the scheme's flexibility based on subjective assessment indicate a stability of nearly 100% in consistent extraction of target images in the standard view, not only for familiar faces but also for unfamiliar faces, when the input face image roughly matches the from view. By combining this scheme for normalizing images into the standard view with an image matching technique for identification, an experimental system for identifying faces among a limited number of subjects was implemented on a commercial engineering workstation High success rates achieved in the identification of front view face images obtained under uncontrolled conditions have objectively confirmed the potential of the scheme for accurate extraction of target images.
We have aimed at constructing a forward dynamics model (FDM) of the human arm in the form of an artificial neural network while recordings of EMG and movement trajectories. We succeeded in: (1) estimating the joint to...
详细信息
We have aimed at constructing a forward dynamics model (FDM) of the human arm in the form of an artificial neural network while recordings of EMG and movement trajectories. We succeeded in: (1) estimating the joint torques under isometric conditions and (2) estimating trajectories from surface EMG signals in the horizontal plane. The human arm has seven degrees of freedom: the shoulder has three, the elbow has one and the wrist has three. Only two degrees of freedom were considered in the previous work. Moreover, the arm was supported horizontally. So, free movement in 3D space is still a necessity. And for 3D movements or posture control, compensation for gravity has to be considered. In this paper, four joint angles, one at the elbow and three at the shoulder were estimated from surface EMG signals of 12 flexor and extensor muscles during posture control in 3D space.
In this paper we first show some of the bifurcation properties of Potts mean-field-theory annealing applied to traveling salesman problems. Due to these bifurcation properties, this approach, in general, produces non-...
详细信息
In this paper we first show some of the bifurcation properties of Potts mean-field-theory annealing applied to traveling salesman problems. Due to these bifurcation properties, this approach, in general, produces non-optimal and non-unique solutions. As an alternative approach, we propose a nonequilibrium version of the Potts spin neural network, called chaotic Ports spin (CPS). CPS has several parameters, and bifurcations over each parameter are investigated. Next, experimental results are shown comparing CPS with several related approaches. CPS is good at obtaining optimal solutions for small-scale problems and semi-optimal solutions for relatively large-scale problems. We also describe a couple of CPS modifications: CPS with a heuristic method and CPS with a ''chaotic annealing'' method. These modified algorithms can produce even better CPS solutions. (C) 1997 Elsevier Science Ltd.
暂无评论