The early stages of diabetic foot represent a critical treatment period, but patients show no obvious symptoms. Upon the development into foot ulcers, a risk of amputation exists for which treatment costs are high. In...
详细信息
The early stages of diabetic foot represent a critical treatment period, but patients show no obvious symptoms. Upon the development into foot ulcers, a risk of amputation exists for which treatment costs are high. In this study, considering the plantar pressure as an important physiological parameter of the foot, we proposed methods to assist the diagnosis of early diabetic foot. Plantar pressure images of early diabetic foot patients were collected and de-noised. An improved automatic regional division algorithm of plantar pressure images was proposed. Laplacian spectrum features were extracted according to the maximum pressure point, pressure center point, and pressure values of the different plantar regions, including plantar shape and tactile features. Finally, based on these data, a support vector classifier was designed and sequential minimal optimization algorithms were used to train the classifier on the plantar pressure data of the left and right foot in 70 subjects to identify early diabetic foot. The results showed that the average recognition rates of the algorithm were high, providing an important reference for the diagnosis of early diabetic foot.
sequentialminimaloptimization (SMO) algorithm is effective in solving large-scale support vector machine (SVM). The existing algorithms all assume that the kernels are positive definite (PD) or positive semi-definit...
详细信息
sequentialminimaloptimization (SMO) algorithm is effective in solving large-scale support vector machine (SVM). The existing algorithms all assume that the kernels are positive definite (PD) or positive semi-definite (PSD) and should meet the Mercer condition. Some kernels, however, such as sigmoid kernel, which originates from neural network and then is extensively used in SVM, are conditionally PD in certain circumstances;in addition, practically, it is often difficult to prove whether a kernel is PD or PSD or not except some well-known kernels. So, the applications of the existing algorithm of SMO are limited. Considering the deficiency of the traditional ones, this algorithm of solving E-SVR with nonpositive semi-definite (non-PSD) kernels is proposed. Different from the existing algorithms which must consider four Lagrange multipliers, the algorithm proposed in this article just need to consider two Lagrange multipliers in the process of implementation. The proposed algorithm simplified the implementation by expanding the original dual programming of E-SVR and solving its KKT conditions, thus being easily applied in solving E-SVR with non-PSD kernels. The presented algorithm is evaluated using five benchmark problems and one reality problem. The results show that E-SVR with non-PSD provides more accurate prediction than that with PD kernel.
Serial numbers identification of RMB (the name of Chinese paper currency) is a nonlinear and high dimensions pattern recognition problem which sample is limited. It is one of many difficulty problems in pattern recogn...
详细信息
ISBN:
(纸本)9781424467129
Serial numbers identification of RMB (the name of Chinese paper currency) is a nonlinear and high dimensions pattern recognition problem which sample is limited. It is one of many difficulty problems in pattern recognition. It also has great research and practical value. This thesis studies the multi-class optimize algorithm in statistical learning theory, analyzes SMOD algorithm and its precondition of serial number recognition. It applies the support vector machine into the serial number's machine recognition of paper currency. It puts forward the theory of serial number identification which based on SVM method, establishes the identification process of identification by SVM. Then we write the number identification algorithm and carrying on simulation test. The experimental results proved that sequentialminimal optimized SVM has fairly low computing load and high precision of recognition. It fully shows the advantages of SVM in solving limited samples, non-linear and high dimension pattern recognition problems. Compared to neural network and fuzzy theory algorithm, its computing load is fairly low. So it can be easily realized with embedded controller.
The sequentialminimaloptimization (SMO) algorithm is a popular algorithm used to solve the Support Vector Machine problem due to its efficiency and ease of implementation. We investigate applying extrapolation metho...
详细信息
ISBN:
(纸本)0780382439
The sequentialminimaloptimization (SMO) algorithm is a popular algorithm used to solve the Support Vector Machine problem due to its efficiency and ease of implementation. We investigate applying extrapolation methods to the SMO update method in order to increase the rate of convergence of this algorithm. We first show that the update method is Newtonian and that extrapolation ensures the update is norm reducing on the objective function. We also note that choosing the working set pair according to some partial order does result in slightly faster speedups in algorithm performance.
暂无评论