classify_class_gmmT_classify_class_gmmClassifyClassGmmClassifyClassGmmclassify_class_gmm (算子)
名称
classify_class_gmmT_classify_class_gmmClassifyClassGmmClassifyClassGmmclassify_class_gmm — 通过高斯混合模型计算特征向量的类。
签名
void ClassifyClassGmm(const HTuple& GMMHandle, const HTuple& Features, const HTuple& Num, HTuple* ClassID, HTuple* ClassProb, HTuple* Density, HTuple* KSigmaProb)
HTuple HClassGmm::ClassifyClassGmm(const HTuple& Features, Hlong Num, HTuple* ClassProb, HTuple* Density, HTuple* KSigmaProb) const
static void HOperatorSet.ClassifyClassGmm(HTuple GMMHandle, HTuple features, HTuple num, out HTuple classID, out HTuple classProb, out HTuple density, out HTuple KSigmaProb)
HTuple HClassGmm.ClassifyClassGmm(HTuple features, int num, out HTuple classProb, out HTuple density, out HTuple KSigmaProb)
def classify_class_gmm(gmmhandle: HHandle, features: Sequence[float], num: int) -> Tuple[Sequence[int], Sequence[float], Sequence[float], Sequence[float]]
def classify_class_gmm_s(gmmhandle: HHandle, features: Sequence[float], num: int) -> Tuple[int, Sequence[float], Sequence[float], Sequence[float]]
描述
classify_class_gmmclassify_class_gmmClassifyClassGmmClassifyClassGmmClassifyClassGmmclassify_class_gmm computes the best NumNumNumNumnumnum classes of
the feature vector FeaturesFeaturesFeaturesFeaturesfeaturesfeatures with the Gaussian Mixture Model
(GMM) GMMHandleGMMHandleGMMHandleGMMHandleGMMHandlegmmhandle and returns the classes in ClassIDClassIDClassIDClassIDclassIDclass_id
and the corresponding probabilities of the classes in
ClassProbClassProbClassProbClassProbclassProbclass_prob. Before calling classify_class_gmmclassify_class_gmmClassifyClassGmmClassifyClassGmmClassifyClassGmmclassify_class_gmm, the
GMM must be trained with train_class_gmmtrain_class_gmmTrainClassGmmTrainClassGmmTrainClassGmmtrain_class_gmm。
classify_class_gmmclassify_class_gmmClassifyClassGmmClassifyClassGmmClassifyClassGmmclassify_class_gmm corresponds to a call to
evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm and an additional step that extracts the
best NumNumNumNumnumnum classes. As described with
evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm, the output values of the GMM can be
interpreted as probabilities of the occurrence of the respective
classes. However, here the posterior probability ClassProbClassProbClassProbClassProbclassProbclass_prob
is further normalized as ClassProbClassProbClassProbClassProbclassProbclass_prob = p(i|x)/p(x)
, where p(i|x) and p(x) are specified with evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm。In most cases it should be sufficient to use NumNumNumNumnumnum =
1 in order to decide whether the
probability of the best class is high enough. In some applications
it may be interesting to also take the second best class into
account (NumNumNumNumnumnum = 2), particularly if it can be
expected that the classes show a significant degree of overlap.
DensityDensityDensityDensitydensitydensity and KSigmaProbKSigmaProbKSigmaProbKSigmaProbKSigmaProbksigma_prob are explained with
evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm。
执行信息
- 多线程类型：可重入（与非独占算子并行运行）。
- 多线程作用域：全局（可从任何线程调用）。
- 未采用并行化处理。
参数
GMMHandleGMMHandleGMMHandleGMMHandleGMMHandlegmmhandle (输入控制) class_gmm → HClassGmm, HTupleHHandleHTupleHtuple (handle) (IntPtr) (HHandle) (handle)
GMM 句柄。
FeaturesFeaturesFeaturesFeaturesfeaturesfeatures (输入控制) real-array → HTupleSequence[float]HTupleHtuple (real) (double) (double) (double)
Feature vector.
NumNumNumNumnumnum (输入控制) integer → HTupleintHTupleHtuple (integer) (int / long) (Hlong) (Hlong)
Number of best classes to determine.
默认值：
1
建议值：
1, 2, 3, 4, 5
ClassIDClassIDClassIDClassIDclassIDclass_id (输出控制) integer(-array) → HTupleSequence[int]HTupleHtuple (integer) (int / long) (Hlong) (Hlong)
Result of classifying the feature vector with
the GMM.
ClassProbClassProbClassProbClassProbclassProbclass_prob (输出控制) real-array → HTupleSequence[float]HTupleHtuple (real) (double) (double) (double)
A-posteriori probability of the classes.
DensityDensityDensityDensitydensitydensity (输出控制) real-array → HTupleSequence[float]HTupleHtuple (real) (double) (double) (double)
Probability density of the feature vector.
KSigmaProbKSigmaProbKSigmaProbKSigmaProbKSigmaProbksigma_prob (输出控制) real-array → HTupleSequence[float]HTupleHtuple (real) (double) (double) (double)
Normalized k-sigma-probability for the feature
vector.
结果
如果参数有效，算子 classify_class_gmmclassify_class_gmmClassifyClassGmmClassifyClassGmmClassifyClassGmmclassify_class_gmm 返回值 2 ( H_MSG_TRUE )。如有必要，则抛出异常。
可能的前趋
train_class_gmmtrain_class_gmmTrainClassGmmTrainClassGmmTrainClassGmmtrain_class_gmm,
read_class_gmmread_class_gmmReadClassGmmReadClassGmmReadClassGmmread_class_gmm
替代
evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm
另见
create_class_gmmcreate_class_gmmCreateClassGmmCreateClassGmmCreateClassGmmcreate_class_gmm
参考文献
Christopher M. Bishop: “Neural Networks for Pattern Recognition”;
Oxford University Press, Oxford; 1995.
Mario A.T. Figueiredo: “Unsupervised Learning of Finite Mixture
Models”; IEEE Transactions on Pattern Analysis and Machine
Intelligence, Vol. 24, No. 3; March 2002.
模块
基础