nvneural::Image Class Reference

Standard implementation for IFileImage. More...

#include <CpuImage.h>

Inheritance diagram for nvneural::Image:

Public Member Functions
std::size_t	channels () const noexcept
	Returns the number of color channels in the image. More...
const std::uint8_t *	data () const noexcept
	This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.
std::uint8_t *	data () noexcept
	Returns a pointer to the raw pixel data in HWC layout. More...
std::size_t	elements () const noexcept
	Returns the number of data elements in the image. More...
NeuralResult	fromFill (std::size_t height, std::size_t width, std::size_t channels, std::uint8_t fillValue) noexcept
	Loads an image by tiling a fill value. More...
NeuralResult	fromRandomValues (std::size_t height, std::size_t width, std::size_t channels) noexcept
	Loads an image by generating normally distributed noise in the range [0, 255]. More...
std::size_t	height () const noexcept
	Returns the height of the image in pixels. More...
NeuralResult	loadFromDescriptors (const std::vector< ImageDescriptor > &descriptors) noexcept
	Loads an image from a chained set of descriptors. More...
NeuralResult	loadFromFile (const std::string &filename) noexcept
	Loads an image from a single bitmap file. More...
NeuralResult	resize (std::size_t height, std::size_t width, std::size_t channels) noexcept
	Clears the image and reserves storage. More...
NeuralResult	saveToFile (const std::string &filename) noexcept
	Saves the resulting image as a file. More...
std::size_t	width () const noexcept
	Returns the width of the image in pixels. More...
Public Member Functions inherited from nvneural::refobj::RefObjectBase< nvneural::refobj::Implements< nvneural::IFileImage >, nvneural::refobj::Implements< nvneural::IImage > >
IRefObject::RefCount	addRef () const noexcept
	Increment the object's reference count. More...
const void *	queryInterface (IRefObject::TypeId interfaceId) const noexcept
	This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.
void *	queryInterface (IRefObject::TypeId interfaceId) noexcept
	Retrieves a new object interface pointer. More...
	RefObjectBase ()
	Default constructor. Logs object creation.
IRefObject::RefCount	release () const noexcept
	Decrements the object's reference count and destroys the object if the reference count reaches zero. More...

Detailed Description

Standard implementation for IFileImage.

Member Function Documentation

channels()

size_t Image::channels ( ) const

noexcept

Returns the number of color channels in the image.

data()

uint8_t * Image::data ( )

noexcept

Returns a pointer to the raw pixel data in HWC layout.

Returns nullptr if the object is not in a valid state.

elements()

size_t Image::elements ( ) const

noexcept

Returns the number of data elements in the image.

This can be calculated as height * width * channels.

fromFill()

NeuralResult Image::fromFill ( std::size_t  height, std::size_t  width, std::size_t  channels, std::uint8_t  fillValue  )

noexcept

Loads an image by tiling a fill value.

Replaces existing data.

Parameters

height	Height of the resulting image
width	Width of the resulting image
channels	Channel count of the resulting image
fillValue	Byte to fill the image with; per-channel fills are not supported

fromRandomValues()

NeuralResult Image::fromRandomValues ( std::size_t  height, std::size_t  width, std::size_t  channels  )

noexcept

Loads an image by generating normally distributed noise in the range [0, 255].

Replaces existing data.

Parameters

height	Height of the resulting image
width	Width of the resulting image
channels	Channel count of the resulting image

height()

size_t Image::height ( ) const

noexcept

Returns the height of the image in pixels.

loadFromDescriptors()

NeuralResult Image::loadFromDescriptors ( const std::vector< ImageDescriptor > &  descriptors )

noexcept

Loads an image from a chained set of descriptors.

Replaces existing data.

At this time, only a single descriptor is supported. Future versions of the framework may support concatenation of image data using sets of descriptors. Loading from npy files is the recommended alternative.

Parameters

descriptors

List of descriptors

loadFromFile()

NeuralResult Image::loadFromFile ( const std::string &  filename )

noexcept

Loads an image from a single bitmap file.

Replaces existing data.

Parameters

filename

File path to load

resize()

NeuralResult Image::resize ( std::size_t  height, std::size_t  width, std::size_t  channels  )

noexcept

Clears the image and reserves storage.

This ensures operations on data() will have an appropriately sized canvas.

saveToFile()

NeuralResult Image::saveToFile ( const std::string &  filename )

noexcept

Saves the resulting image as a file.

Parameters

filename

File path to save

width()

size_t Image::width ( ) const

noexcept

Returns the width of the image in pixels.

---

The documentation for this class was generated from the following files:

• Helpers/Inc/CpuImage.h
• Helpers/Src/CpuImage.cpp