Getting started with VLAD

本篇博客介绍了如何使用C API计算一副图像的VLAD，更多细节请看http://www.vlfeat.org/api/vlad-fundamentals.html（VLAD的定义和计算）

函数vl_vlad_encode能够获得一系列的VLAD特征，数据类型可以是float或者double。它需要一个视觉词典，比如说从K-means聚类中得到的，进一步，词典元素的特征分配必须是预处理的，比如通过KD-trees和forests。

void vl_vlad_encode	(	void *	enc,
		vl_type	dataType,
		void const *	means,
		vl_size	dimension,
		vl_size	numClusters,
		void const *	data,
		vl_size	numData,
		void const *	assignments,
		int	flags
	)

Parameters

enc	output VLAD encoding (out).
dataType	the type of the input data (VL_TYPE_DOUBLE or VL_TYPE_FLOAT).
numData	number of data vectors to encode.
means	cluster means.
numClusters	number of clusters.
data	the data vectors to encode.
dimension	dimensionality of the data.
assignments	data to cluster soft assignments.
flags	options.

在后面的例子中，词典时首先通过Kmeans聚类产生的，然后是点被分配到它的相关临近的词典中， 替代了原始的VLAD特征。最后编码存储在enc变量中。

vl_uint32 * indexes;

float * assignments;

float * enc

int i;

// create a KMeans object and run clustering to get vocabulary words (centers)

kmeans = vl_kmeans_new (VLDistanceL2, VL_TYPE_FLOAT) ;

vl_kmeans_cluster (kmeans,

data,

dimension,

numData,

numCenters) ;

// find nearest cliuster centers for the data that should be encoded

indexes = vl_malloc( sizeof( vl_uint32) * numDataToEncode);

vl_kmeans_quantize(kmeans,indexes,dataToEncode,numDataToEncode);

// convert indexes array to assignments array,

// which can be processed by vl_vlad_encode

assignments = vl_malloc( sizeof( float) * numDataToEncode * numCenters);

memset(assignments, 0, sizeof( float) * numDataToEncode * numCenters);

for(i = 0; i < numDataToEncode; i++) 

{

assignments[i * numCenters + indexes[i]] = 1.;

}

// allocate space for vlad encoding

enc = vl_malloc( sizeof(TYPE) * dimension * numCenters);

// do the encoding job

vl_vlad_encode (enc, VL_F_TYPE,

vl_kmeans_get_centers(kmeans), dimension, numCenters,

data, numData,

assignments,

0) ;