Template Class TensorDataAccessStridedImageImpl

• Defined in File TensorDataAccess.hpp

Inheritance Relationships

Base Type

• public nvcv::detail::TensorDataAccessStridedImpl< ShapeInfo > (Template Class TensorDataAccessStridedImpl)

Derived Types

• public nvcv::detail::TensorDataAccessStridedImagePlanarImpl< TensorShapeInfoImagePlanar > (Template Class TensorDataAccessStridedImagePlanarImpl)

• public nvcv::detail::TensorDataAccessStridedImagePlanarImpl< ShapeInfo > (Template Class TensorDataAccessStridedImagePlanarImpl)

Class Documentation

template<typename ShapeInfo>
class TensorDataAccessStridedImageImpl : public nvcv::detail::TensorDataAccessStridedImpl<ShapeInfo>

Provides specialized access methods for strided tensor data representing images.

This class is an extension of TensorDataAccessStridedImpl and offers specific utilities for accessing the data in image tensors using a strided memory layout. It provides methods to retrieve the number of columns, rows, channels, and other image-specific properties. Furthermore, it provides utility methods to compute strides and access specific rows, channels, etc.

Template Parameters:

ShapeInfo – The type that contains shape information for the image tensor.

Subclassed by nvcv::detail::TensorDataAccessStridedImagePlanarImpl< TensorShapeInfoImagePlanar >, nvcv::detail::TensorDataAccessStridedImagePlanarImpl< ShapeInfo >

Public Functions

inline TensorDataAccessStridedImageImpl(const TensorDataStrided &tdata, const ShapeInfo &infoShape)

Constructor that initializes the object with the given tensor data and shape information.

Parameters:

• tdata – The strided tensor data.

• infoShape – The shape information of the tensor.

inline int32_t numCols() const

Returns the number of columns in the image tensor.

Returns:

The number of columns.

inline int32_t numRows() const

Returns the number of rows in the image tensor.

Returns:

The number of rows.

inline int32_t numChannels() const

Returns the number of channels in the image tensor.

Returns:

The number of channels.

inline Size2D size() const

Returns the size (width and height) of the image tensor.

Returns:

The size of the image.

inline int64_t chStride() const

Computes the stride for the channel dimension.

Returns:

The channel stride, or 0 if the channel dimension is not present.

inline int64_t colStride() const

Computes the stride for the column (or width) dimension.

Returns:

The column stride, or 0 if the width dimension is not present.

inline int64_t rowStride() const

Computes the stride for the row (or height) dimension.

Returns:

The row stride, or 0 if the height dimension is not present.

inline int64_t depthStride() const

Computes the stride for the depth dimension.

Returns:

The depth stride, or 0 if the depth dimension is not present.

inline Byte *rowData(int y) const

Retrieves a pointer to the data of a specific row.

Parameters:

y – The row index.

Returns:

A pointer to the row’s data.

inline Byte *rowData(int y, Byte *base) const

Retrieves a pointer to the data of a specific row, relative to a given base pointer.

Parameters:

• y – The row index.

• base – The base pointer from which to compute the row’s data address.

Returns:

A pointer to the row’s data.

inline Byte *chData(int c) const

Retrieves a pointer to the data of a specific channel.

Parameters:

c – The channel index.

Returns:

A pointer to the channel’s data.

inline Byte *chData(int c, Byte *base) const

Retrieves a pointer to the data of a specific channel, relative to a given base pointer.

Parameters:

• c – The channel index.

• base – The base pointer from which to compute the channel’s data address.

Returns:

A pointer to the channel’s data.

Protected Functions

inline TensorDataAccessStridedImageImpl(const TensorDataAccessStridedImageImpl &that, const ShapeInfo &infoShape)