curfil | |
utils | |
Timer | Used to get seconds or milliseconds passed for profiling purposes |
Average | A simple class used for averaging |
Profile | Used to profile different stages of training and prediction |
RandomTreeExport | Helper class to export a random tree or random forest to disk in compressed (gzip) JSON format |
Result | Class that stores the results of a hyperopt run |
HyperoptClient | Client that does a hyperopt parameter search |
RGBColor | A tuple of three 8-bit uints that represent a RGB color value |
Depth | Wrapper class that represent a depth value as it occurs in RGB-D images |
RGBDImage | An RGB-D image that contains four channels for the RGB color and the depth |
LabelImage | A labelling that usually belongs to a RGBImage |
LabeledRGBDImage | A tuple of a RGBD image and an according labeling |
RandomTreeImport | Helper class to import a random tree or random forest from disk in compressed (gzip) JSON format |
ConfusionMatrix | Helper class to maintain a n×n confusion matrix |
RandomForestImage | A random forest for RGB-D images |
SplitFunction | Class used to get the splitting score for a given feature and threshold |
TrainingConfiguration | Parameters used for training, most of which are provided by the user |
RandomTree | A sub-tree with histograms and pointers to the parent and children |
Sampler | A uniform distribution sampler |
ReservoirSampler | Stores samples (images or pixel instances), if max size is reached, a sample replaces another chosen at random |
RandomSource | Used to get a uniform sampler after incrementing the seed |
RandomTreeTrain | Class that does the actual training of a tree |
XY | A simple 2D tuple |
PixelInstance | Represents a single-pixel sample in a RGB-D image |
ImageFeatureFunction | Parametrized visual image feature |
ImageFeaturesAndThresholds | Helper class to store a list of features and a list of threshold per feature in a compact mannersuch that it can be transferred between CPU and GPU |
Samples | Helper class to store a list of pixel instances in a compact mannersuch that it can be transferred between CPU and GPU |
ImageFeatureEvaluation | Helper class for the four phases in the cost-intensive best-split evaluation during random forest training |
RandomTreeImage | A random tree in a random forest, for RGB-D images |
TreeNodes | Helper class to map random forest data to texture cache on GPU |
DeviceCache | Abstract base class that implements a simple LRU cache on GPU |
ImageCache | A simple LRU cache of RGB-D images on GPU |
TreeCache | A simple LRU cache of trees on GPU |
TreeNodeData | Helper class for the unit test |
InformationGainScore | Information Gain calculation using Shannon entropy of the parent and child nodes |
NormalizedInformationGainScore | Information gain normalized by the sum of the classification and split entropies see formula (11) in [Wehenkel1991] and appendix A in [Geurts2006] |
cuv | |
allocator | |
default_allocator | Allocator allows allocation, deallocation and copying depending on memory_space_type |
cuda_allocator | Allocator that uses cudaMallocHost for allocations in host_memory_space |
pooled_cuda_allocator | Allocator that naively pools device and host memory |
column_major | Tag for column major matrices |
row_major | Tag for row major matrices |
linear_memory_tag | Tag for linear memory |
pitched_memory_tag | Tag for pitched memory |
memory | Simply keeps a pointer and deallocates it when destroyed |
linear_memory | Contiguous memory |
pitched_memory | 2D non-contiguous ("pitched") memory |
FalseType | Defines "False" |
TrueType | Defines "True" |
IsSame | Checks whether two types are equal |
IsSame< T, T > | |
IsDifferent | Checks whether two types are different |
IsDifferent< T, T > | |
unconst | Remove "const" from a type |
unconst< const T > | |
If | Switch result depending on Condition |
If< false, Then, Else > | |
EnableIfC | Enable-if controlled creation of SFINAE conditions |
EnableIfC< false, T > | |
EnableIf | |
DisableIf | |
ndarray_info | Infos about shape and stride on host and in the ndarray data space |
ndarray | N-dimensional array on GPU or CPU |
ndarray_view | Primarily used as result of ndarray::operator[] |
switch_value_type | Create a ndarray type with the same template parameters, but with switched value type |
switch_memory_layout_type | Create a ndarray type with the same template parameters, but with switched memory_layout_type |
switch_memory_space_type | Create a ndarray type with the same template parameters, but with switched memory_space_type |
reference | This objects acts like a reference to the object stored at the wrapped pointer |
host_memory_space | Tag for host memory |
dev_memory_space | Tag for device memory |