last update 20 Sep 2009 |
Classes | |
class | cvr::distanceType< T > |
This class defines the appropriate return types when calculating the distance between two values of type T. More... | |
Typedefs | |
typedef unsigned _CVR_BYTE | cvr::ubyte |
typedef signed _CVR_BYTE | cvr::byte |
typedef unsigned _CVR_WORD16 | cvr::uint16 |
typedef signed _CVR_WORD16 | cvr::int16 |
typedef unsigned _CVR_WORD32 | cvr::uint32 |
typedef signed _CVR_WORD32 | cvr::int32 |
typedef float | cvr::sreal |
typedef double | cvr::dreal |
Enumerations | |
enum | cvr::eBoundaryType { cvr::Zero, cvr::Mirror, cvr::Periodic, cvr::Constant, cvr::NoBoundary } |
enum | cvr::eConstantReference { cvr::VariableReference, cvr::ConstantReference } |
enum | cvr::eResizeType { cvr::AllocateOnly, cvr::Copy, cvr::Init, cvr::CopyAndInit } |
Variables | |
__extension__ typedef unsigned _CVR_WORD64 | cvr::uint64 |
__extension__ typedef signed _CVR_WORD64 | cvr::int64 |
typedef signed _CVR_BYTE cvr::byte |
8-bit signed type (-128.
.127)
typedef double cvr::dreal |
64-bit floating point type
typedef signed _CVR_WORD16 cvr::int16 |
16-bit signed type (-32768.
.32767)
typedef signed _CVR_WORD32 cvr::int32 |
32-bit signed type (-(2^31).
.(2^31-1))
typedef float cvr::sreal |
32-bit floating point type
typedef unsigned _CVR_BYTE cvr::ubyte |
8-bit unsigned type (0.
.255)
typedef unsigned _CVR_WORD16 cvr::uint16 |
16-bit unsigned type (0.
.65535)
typedef unsigned _CVR_WORD32 cvr::uint32 |
32-bit unsigned type (0.
.(2^32)-1)
enum cvr::eBoundaryType |
boundary condition type.
This type indicates how to consider/interpret the boundaries of an image or matrix.
For this type the methods read(ioHandler&, eBoundaryType&) and write(ioHandler&, const eBoundaryType&) are implemented. Note that template functions read(ioHandler&, const std::string&, T&) and write(ioHandler&, const std::string&, const T&) exist. (see Developer's Guide for further information).
Zero | The boundary does not exist (only 0-value). |
Mirror | The signal/image/vector is mirrored. |
Periodic | A periodic continuation of the signal is assumed. |
Constant | The last value of the border will be assumed. |
NoBoundary |
The boundary will be ignored.
The effect is an unitialized border in those parts of the vector, matrix or image, where a complete filter kernel cannot be applied due to its size. Please note that the border will be left UNinitialized. It is up to you to set some value at this border, which will have a maximum height or width of the size of the filter kernel minus one. The reason for this option to exist is that some algorithms will ignore this border anyway and all other border types will spend some time calculating it |
The constant reference type is used to indicate if the memory allocated by vectors, matrices and tensors can be modified or not.
It is used to avoid that the subtensors (or vectors) of tensors (or matrices) modified the memory they use, because it is allocated outside.
enum cvr::eResizeType |
The resize type defines the behaviour of resize member functions found in many containers and other data structures.
In most cases the default is Copy, to be compliant with the STL. Different functions: allocate() and assign() are shortcuts for resize with OnlyAllocate or Init.
For this type the methods read(ioHandler&, eResizeType&) and write(ioHandler&, const eResizeType&) are implemented. Note that template functions read(ioHandler&, const std::string&, T&) and write(ioHandler&, const std::string&, const T&) exist. (see Developer's Guide for further information).
AllocateOnly | Memory is allocated only, old values are not copied to the new container and new fields are not initialized. |
Copy | Old values are copied to the resized container but possibly existing new elements are not initialized. |
Init |
The resized container is completely filled with the given init value.
|
CopyAndInit | The old data is copied to the resized container and new elements (if existent) are initialized with the init value. |
__extension__ typedef signed _CVR_WORD64 cvr::int64 |
64-bit signed type (-(2^63).
.(2^63-1))
__extension__ typedef unsigned _CVR_WORD64 cvr::uint64 |
64-bit unsigned type (0.
.(2^64)-1)