This paper proposes a semi-automatic segmentation algorithm using a double labeling method. Semi-automatic segmentation consists of intra-frame segmentation and inter-frame segmentation. For intra-frame segmentation, ...
详细信息
This paper proposes a semi-automatic segmentation algorithm using a double labeling method. Semi-automatic segmentation consists of intra-frame segmentation and inter-frame segmentation. For intra-frame segmentation, we obtain a globally labeled mask from the boundary information provided by the user and a locally labeled mask from the intensity information of the image itself. By merging the globally and locally labeled masks, we complete the intra-frame segmentation operation. Inter-frame segmentation is accomplished by tracking the video objects (VOs) that are defined from intra-frame segmentation. Semi-automatic segmentation using the double labeling method resolves the object correspondence problem, which is inherent in automatic segmentation.
An approach to fault isolation of multiplicative faults in linear systems is presented. After modeling the faults as multiplicative ones, a procedure for the design of structured residuals is proposed. The design is b...
详细信息
An approach to fault isolation of multiplicative faults in linear systems is presented. After modeling the faults as multiplicative ones, a procedure for the design of structured residuals is proposed. The design is based on a modified version of a nonlinear disturbance decoupling algorithm used for nonlinear systems. Nonlinear techniques are required because of the time variant fault distribution matrix resulting of the multiplicative fault modeling. The contributions are the study of fault isolability, the relaxing of existence condition with respect to early results and the develop of an algorithm for the fault decoupling.
The complexity of modern anesthesia procedures requires the development of decision support systems functioning in a smart-alarm capacity. Based on computerized anesthesia records containing hemodynamic data (heart ra...
详细信息
The complexity of modern anesthesia procedures requires the development of decision support systems functioning in a smart-alarm capacity. Based on computerized anesthesia records containing hemodynamic data (heart rate, mean arterial pressure and systolic arterial pressure) and assessments made by experienced anesthesiologists reviewing printed anesthesia records, we have developed rule-based computer algorithms to detect critical conditions during surgery, such as inadequate (light) anesthesia (LA) or unstable blood pressure (lability). Our analysis indicated that a /spl ges/12% change in mean arterial blood pressure (MAP), compared with the median value of MAP over the preceding 10 minute interval, may be chosen as the criterion for detecting LA, with a sensitivity of 96% and a specificity of 91%. The best agreement between human and computer ratings of blood pressure lability (correlation coefficient 0.78) was achieved when we used the absolute value of the fractional change of the mean arterial pressure (|FCM|) between one 2-min epoch and the next 2-min epoch. We developed rule-based computer algorithms to defect critical conditions during surgery (light anesthesia or unstable blood pressure), based on computerized anesthesia records containing hemodynamic data (heart rate, mean arterial pressure and systolic arterial pressure).
An adaptive algorithm is considered in on-line learning of probability functions, which infers a distribution underlying observed data x/sub 1/, x/sub 2/, ..., x/sub N/. The algorithm is based on how we can detect the...
详细信息
An adaptive algorithm is considered in on-line learning of probability functions, which infers a distribution underlying observed data x/sub 1/, x/sub 2/, ..., x/sub N/. The algorithm is based on how we can detect the change of a source function in an unsupervised learning scheme. This is an extension of an optimal on-line learning algorithm of probability distributions, which is derived from the field theoretical point of view. Since we learn not parameters of a model but probability functions themselves, the algorithm has the advantage that it requires no a priori knowledge of a model.
This paper proposes a new method for the extraction of background objects, especially the extraction of randomly moving objects such as curtains and plants. For this purpose we use an optical flow method combined with...
详细信息
This paper proposes a new method for the extraction of background objects, especially the extraction of randomly moving objects such as curtains and plants. For this purpose we use an optical flow method combined with a split-and-merge algorithm to avoid the mentioned background effects. The split-and-merge algorithm is based on the variance of regions and the optical flow method is implemented conform to the proposition of Horn and Schunck (1981). To further increase the robustness and utility in surveillance circumstances, illumination invariance has to be provided by the algorithm. Hence we combined the mentioned methods with an illumination-invariant change detector. Tests have been performed on indoor and outdoor surveillance sequences. Especially in indoor sequences the proposed method provided a very robust tool against illumination changes, reflections, shadows and randomly moving objects like curtains and plants. Nevertheless in all circumstances it detected the moving 'semantic' objects such as persons.
The various algorithms of definition of a spatial rule of border of two environments with different density are investigated for reconstruction of distribution of density on backscatter radiation (in Compton tomograph...
详细信息
The various algorithms of definition of a spatial rule of border of two environments with different density are investigated for reconstruction of distribution of density on backscatter radiation (in Compton tomography). The real missing beams, range of energy 10-1000 keV and various materials are considered. The algorithm based joint processing of functions, N(x) change of number of quantums N in the detector, is offered at change of coordinate x of disseminating volume and its by derivative /spl part/N//spl part/x. The algorithm is steady for a wide class of materials and provides an error of definition of a rule of border of two environments no more than /spl plusmn/0.1 mm.
To implement speaker verification (SV) technology for real-world applications with a large user population, the system cost becomes an important issue. One needs a fast algorithm which can support more users in a cent...
详细信息
ISBN:
(纸本)0780350413
To implement speaker verification (SV) technology for real-world applications with a large user population, the system cost becomes an important issue. One needs a fast algorithm which can support more users in a central telephone switch given the limited hardware, or can reduce the hardware requirement on a wireless handset. Li (see Proc. Int'l. Conf. on Spoken Language Processing, Sydney, 1998) proposed a fast, sequential decoding algorithm for left-to-right HMM. The algorithm is based on a sequential detection scheme which is asymptotically optimal in the sense of detecting a possible change in distribution as reliably and quickly as possible. In this paper, the algorithm is evaluated in a fixed-phrase SV system on a database with 23,578 utterances recorded from 100 speakers. The experimental results show that the decoding speed of the proposed algorithm is about 7 to 10 times faster than the Viterbi algorithm while the accuracy is in an acceptable level. The results indicate that the proposed algorithm can also be applied to speaker identification, utterance verification, audio segmentation, voice/silence detection and many other applications.
We show a technique for consistent dynamic exchange of different aspects of implementations of abstract data types, in particular data structure representations and algorithms. Previous work showed how one single aspe...
详细信息
We show a technique for consistent dynamic exchange of different aspects of implementations of abstract data types, in particular data structure representations and algorithms. Previous work showed how one single aspect's implementation can be exchanged dynamically without running into robustness problems if other aspects are guaranteed to be fixed. The present work generalizes this idea in the sense that all aspects' implementations may be exchanged arbitrarily at run time where robustness is preserved. We use finite state machines where each state represents a combination of the different implementations and each transition an exchange of exactly one aspect's implementation. We use these state machines to check for the robustness of a class within a certain application context.
The prediction of non-stationary dynamical systems may be performed by identifying appropriate sub-dynamics and an early detection of mode changes. We present a framework which unifies the mixtures of experts approach...
详细信息
The prediction of non-stationary dynamical systems may be performed by identifying appropriate sub-dynamics and an early detection of mode changes. We present a framework which unifies the mixtures of experts approach and a generalized hidden Markov model with an input-dependent transition matrix: the hidden Markov mixtures of experts (HMME). The gating procedure incorporates state memory, information about the current location in phase space, and the previous prediction performance. The experts and the hidden Markov gating model are simultaneously trained by an EM algorithm that maximizes the likelihood during an annealing procedure. The HMME architecture allows for a fast online detection of mode changes: change points are detected as soon as the incoming input data stream contains sufficient information to indicate a change in the dynamics.
暂无评论