We introduce a new graph cut for clustering which we call the Information Cut. It is derived using Parzen windowing to estimate an information theoretic distance measure between probability density functions. We propose to optimize the Information Cut using a gradient descent-based approach. Our algorithm has several advantages compared to many other graph-based methods in terms of determining an appropriate affinity measure, computational complexity, memory requirements and coping with different data scales. We show that our method may produce clustering and image segmentation results comparable or better than the state-of-the art graph-based methods. 相似文献
The simultaneous synthesis and densification of dense TiC–TiB2 composite has been investigated starting from Ti, B4C, and C as reactants, and using the spark plasma sintering (SPS) technique. The optimal conditions for complete conversion of the reactants to the composite were determined for different applied DC current levels. A kinetic investigation performed allows us to conclude that solid-state diffusion is the mechanism governing the synthesis process. It is seen that TiC is the first phase formed, while TiB2 formation occurs afterward. Two intermediate boride phases, i.e. TiB and Ti3B4, are also formed but, as the SPS holding time was augmented, they were gradually and completely converted to TiB2. Moreover, it is found that in order to reach relatively high dense products, an electric current needs to be applied for time intervals longer than those required for obtaining complete conversion. A pure dense product (relative density ∼98%) was obtained when an electric pulsed current of 1100 A and a mechanical pressure of 20 MPa were applied for about 4 min. 相似文献
Studying the changes of shape is a common concern in many scientific fields. We address here two problems: (1) quantifying the deformation between two given shapes and (2) transporting this deformation to morph a third shape. These operations can be done with or without point correspondence, depending on the availability of a surface matching algorithm, and on the type of mathematical procedure adopted. In computer vision, the re-targeting of emotions mapped on faces is a common application. We contrast here four different methods used for transporting the deformation toward a target once it was estimated upon the matching of two shapes. These methods come from very different fields such as computational anatomy, computer vision and biology. We used the large diffeomorphic deformation metric mapping and thin plate spline, in order to estimate deformations in a deformational trajectory of a human face experiencing different emotions. Then we use naive transport (NT), linear shift (LS), direct transport (DT) and fanning scheme (FS) to transport the estimated deformations toward four alien faces constituted by 240 homologous points and identifying a triangulation structure of 416 triangles. We used both local and global criteria for evaluating the performance of the 4 methods, e.g., the maintenance of the original deformation. We found DT, LS and FS very effective in recovering the original deformation while NT fails under several aspects in transporting the shape change. As the best method may differ depending on the application, we recommend carefully testing different methods in order to choose the best one for any specific application.
Distributed and Parallel Databases - Given two datasets of points (called Query and Training), the Group (K) Nearest-Neighbor (GKNN) query retrieves (K) points of the Training with the smallest sum... 相似文献
Neural Computing and Applications - In this paper, we present the novel Deep-MEG approach in which image-based representations of magnetoencephalography (MEG) data are combined with ensemble... 相似文献
Recent advances in Wireless Mesh Networks (WMNs) have overcome the drawbacks of traditional wired and ad-hoc networks and now they are seen as a means of allowing last mile communications with quality level assurance in Future Multimedia Systems. However, new routing schemes are needed to provide end-to-end Quality of Service (QoS) and Quality of Experience (QoE) support for delay/loss/jitter-sensitive multimedia applications. The well-known OLSR (Optimized Link State Routing) protocol with ETX (Expected Transmission Count) metric brings many benefits to the path selection process, but has a drawback with regard to queue availability management, which reduces the system performance. This problem is caused when OLSR-EXT control messages are exchanged and the queues of mesh routers along the end-to-end communication path are overloaded. As a result, multimedia-related packets will suffer from loss/delay/jitter and the overall system performance will decrease. This paper proposes the Optimized Link State Routing-Fuzzy ETX Queue (OLSR-FEQ) protocol to overcome the limitations of OLSR-ETX regarding queue availability, QoS and QoE assurance. OLSR-FEQ optimizes network and user-based parameters by coordinating queue availability, QoS and fuzzy issues in the routing decision process as a way of allocating the best paths for multimedia applications. Performance evaluations were carried out with the Network Simulator (NS-2.34) to show the benefits of the proposed solution when compared with existing routing schemes, namely OLSR-ETX, OLSR-FLC, OLSR-MD and HWMP (IEEE 802.11s standard), regarding QoS (unsuccessful packet delivery and throughput) and QoE (PSNR, SSIM, VQM and MOS) parameters. 相似文献
This work is part of a project aimed to develop automotive real-time observers based on detailed nonlinear multibody models
and the extended Kalman filter (EKF). In previous works, a four-bar mechanism was studied to get insight into the problem.
Regarding the formulation of the equations of motion, it was concluded that the state-space reduction method known as matrix-R
is the most suitable one for this application. Regarding the sensors, it was shown that better stability, accuracy and efficiency
are obtained as the sensored magnitude is a lower derivative and when it is a generalized coordinate of the problem. In the
present work, the automotive problem has been addressed, through the selection of a Volkswagen Passat as a case-study. A model
of the car containing fifteen degrees of freedom has been developed. The observer algorithm that combines the equations of
motion and the integrator has been reformulated so that duplication of the problem size is avoided, in order to improve efficiency.
A maneuver of acceleration from rest and double lane change has been defined, and tests have been run for the “prototype,”
the “model” and the “observer,” all the three computational, with the model having 100 kg more than the prototype. Results
have shown that good convergence is obtained for position level sensors, but the computational cost is high, still far from
real-time performance. 相似文献
The recent popularity of digital cameras has posed a new problem: how to efficiently store and retrieve the very large number
of digital photos captured and chaotically stored in multiple locations without any annotation. This paper proposes an infrastructure,
called PhotoGeo, which aims at helping users with the people photo annotation, event photo annotation, storage and retrieval
of personal digital photos. To achieve the desired objective, PhotoGeo uses new algorithms that make it possible to annotate
photos with the key metadata to facilitate their retrieval, such as: the people who were shown in the photo (who); where it was captured (where); the date and time of capture (when); and the event that was captured. The paper concludes with a detailed evaluation of these algorithms. 相似文献
In healthcare facilities most of the daily activities require strict coordination between clinicians, who often operate under
heavy workloads and minimal workforce conditions in environments filled with increasingly complex technology. Ubiquitous Computing
applications constitute a suitable solution for both reducing medical costs and improving patient safety by better supporting
clinical processes. In this study we introduce an intelligent infrastructure for smart hospitals which implements basic services
to optimize medical staff/patient interactions and grants ubiquitous and transparent access to clinical data stored in standard
clinical databases. This infrastructure relies on the integration of Radio Frequency IDentification (RFID) and photosensor
technologies to identify, locate and track clinicians and patients equipped with mobile devices and wearable RFID tags. 相似文献