Image Resizer

Change dimensions of image in pixel stream

Since R2023a

• 

Libraries:
Vision HDL Toolbox / Geometric Transforms

Description

The Image Resizer block downscales grayscale input images. The horizontal and vertical scale factors are independent and you can optionally specify input and output sizes in real time. The block provides an optional antialiasing filter, and you can choose nearest-neighbor or bilinear interpolation.

You can use this block for image resize operations such as:

• Converting images to a uniform size for input to deep learning algorithms

• Modifying disparate source videos for processing, resource, or display requirements

Examples

Runtime Image Resizing

Configure and use the Image Resizer block with variable input and output image sizes.

• Open Script

Change Image Size

Downsample a multicomponent image by using the Image Resizer block. The example also shows how to implement custom bicubic and Lanczos-2 interpolation algorithms for FPGAs by using basic Simulink blocks.

• Open Model

Ports

This block uses a streaming pixel interface with a pixelcontrol bus for frame control signals. All Vision HDL Toolbox™ blocks use the same streaming interface. The block accepts and returns a scalar pixel value and a bus that contains five control signals. The control signals indicate the validity of each pixel and its location in the frame. To convert a frame (pixel matrix) into a serial pixel stream and control signals, use the Frame To Pixels block. For a full description of the interface, see Streaming Pixel Interface.

Input

pixel — Input pixel stream

scalar

Single pixel, specified as a scalar value.

The software supports double and single data types for simulation, but not for HDL code generation.

Data Types: uint8 | uint16 | fixdt(0,WL,0), WL between 8 and 16 | double | single

ctrl — Control signals associated with pixel stream

pixelcontrol bus

The pixelcontrol bus contains five signals. The signals describe the validity of the pixel and its location in the frame. For more information, see Pixel Control Bus.

Data Types: bus

inputSize — Dimensions of active frame in input stream

[ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the input frame, specified as a row vector of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

The software supports double and single data types for simulation, but not for HDL code generation.

Dependencies

To enable this port, set the Input size source parameter to Input port.

If you enable both inputSize and outputSize ports, and those inputs use a fixed point type, both ports must use the same fixed point type.

Data Types: single | double | uint8 | uint16 | fixdt(0,WL,0), WL up to 16

outputSize — Dimensions of active frame in output stream

[ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the output frame, specified as a row vector of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

The output dimensions must be less than or equal to the input dimensions.

The software supports double and single data types for simulation, but not for HDL code generation.

Dependencies

This port appears when you set the Output size source parameter to Input port.

If you enable both inputSize and outputSize ports, and those inputs use a fixed point type, both ports must use the same fixed point type.

Data Types: single | double | uint8 | uint16 | fixdt(0,WL,0), WL up to 16

Output

pixel — Output pixel stream

scalar

Configure the data type of the output pixel by using the Output parameter on the Data Type tab.

Data Types: uint8 | uint16 | fixdt(0,WL,0), WL between 8 and 16 | double | single

ctrl — Control signals associated with pixel stream

pixelcontrol bus

The pixelcontrol bus contains five signals. The signals describe the validity of the pixel and its location in the frame. For more information, see Pixel Control Bus.

Data Types: bus

Parameters

Main

Input height and width — Dimensions of active frame in input stream

[1080,1920] (default) | [ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the input frame, specified as a row vector of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

Dependencies

This parameter appears when you set the Input size source parameter to Property.

Input size source — Source of input active frame size

Property (default) | Input port

Source of the input active frame size, specified as one of these options:

• Property — Specify the input size by using the Input height and width parameter.

• Input port — Specify the input size by using the inputSize input port. The block samples the inputSize and outputSize ports when vStart is true in the input control bus.

Maximum input height and width — Active frame size that accommodates the largest anticipated input size

[1080,1920] (default) | [ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the largest anticipated input frame, specified as a row vector of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

Dependencies

This parameter appears when you set the Input size source parameter to Input port.

Output height and width — Dimensions of active frame in output stream

[720,1280] (default) | [ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the output frame, specified as a matrix of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

The output dimensions must be less than or equal to the input dimensions.

Dependencies

This parameter appears when you set the Output size source parameter to Property.

Output size source — Source of input active frame size

Property (default) | Input port

Source of the output active frame size, specified as one of these options:

• Property — Specify the output size by using the Output height and width parameter.

• Input port — Specify the output size by using the outputSize input port. The block samples the inputSize and outputSize ports when vStart is true in the input control bus.

Maximum output height and width — Active frame size that accommodates the largest anticipated output size

[720, 1280] (default) | [ActiveVideoLines,ActivePixelsPerLine]

Dimensions of the largest anticipated output frame, specified as a row vector of the form [ActiveVideoLines,ActivePixelsPerLine]. The dimensions must be integers greater than or equal to [16,16] and less than or equal to [4320,8192].

Dependencies

This parameter appears when you set the Output size source parameter to Input port.

Insert horizontal blanking — Add invalid cycles between active lines

off (default) | on

The horizontal blanking interval is the period of inactive cycles between the end of one line and the beginning of the next line. When you select this option, the block includes a buffer to store the output pixels while inserting extra horizontal blanking cycles between the active lines of the output video stream. Set the number of blanking cycles in the Minimum output horizontal blanking parameter. For information on when to use horizontal blanking, see Configure Blanking Intervals.

Minimum output horizontal blanking — Number of inactive cycles to insert between output lines

370 (default)

The horizontal blanking interval is the period of inactive cycles between the end of one line and the beginning of the next line. If your downstream processing requires a certain blanking interval, specify that interval in this parameter. This parameter value must be an integer between 1 and 1195. The block includes a buffer that stores the output pixels until the minimum number of blanking cycles is reached.

The recommended minimum horizontal blanking interval for Vision HDL Toolbox library blocks that use a line buffer is 2×Kw when using padding or 12 cycles when you set the Padding method parameter to None. For more information on blanking requirements for image processing systems, see Configure Blanking Intervals.

Dependencies

To enable this parameter, select Insert horizontal blanking.

Interpolation method — Method of interpolating pixel values

Bilinear (default) | Nearest Neighbor

Method of interpolating pixel values in horizontal and vertical directions, specified as one of these types:

• Bilinear — The block implements a two-tap filter in each direction, where the taps are the current and previous pixel values.

• Nearest Neighbor — The object implements muxes to select the closest of the current pixel value or the previous pixel value. This interpolation method uses the fewest hardware resources, at the tradeoff of coarser interpolation accuracy.

For more information about interpolation methods, see the Change Image Size example.

Perform antialiasing on downscale — Apply antialiasing filter before resizing

on (default) | off

When you set this property to true, the object applies an antialiasing filter to the input image before the resize operation. The antialiasing filter is a separable 5-by-5 filter. The coefficients are [1 4 6 4 1]/16.

Data Types

Rounding mode — Rounding method for internal fixed-point calculations

Floor (default) | Ceiling | Convergent | Nearest | Round | Zero

Specify a rounding method for internal fixed-point calculations.

Saturate on integer overflow — Overflow action for internal fixed-point calculations

off (default) | on

When you clear this parameter, fixed-point and integer values wrap around to zero when the value overflows what is representable with that data type. When you select this parameter, the value saturates at the maximum representable value.

Coefficients — Interpolation filter coefficients data type

fixdt(0,16) (default) | <data type expression>

Specify the data type of the filter coefficients as a data type expression in the form fixdt(0,WL), where the type is unsigned and WL is a word length in the range 2 to 17. The block computes the binary point scaling to accommodate the interpolation coefficients. When converting the coefficients to the specified data type, the block uses the settings of the Rounding method and Overflow action parameters.

The interpolation filter also limits the internal pixel data types to 24 bits wide so that the filter maps to DSP blocks on your FPGA.

Output — Output data type

Inherit: Same as first input (default) | fixdt(0,16,0) | <data type expression>

Specify the data type of the output pixels as Inherit: Same as first input or as a data type expression in the form fixdt(0,WL,0), where the type is unsigned, and WL is a word length in the range 8 to 16.

• Inherit: Same as first input — Sets the data type of the output pixels to match the data type of the input pixel port.

• <data type expression> — Sets the data type of the output pixels to this data type.

When converting the pixel values to the specified data type, the block uses the settings of the Rounding method and Overflow action parameters.

Algorithms

The resize architecture consists of a scale calculator, buffer, interpolator, and optional antialiasing filter and output buffer. The scale calculator converts the specified input and output sizes into accumulators that control the selection of pixels from the buffer. The input and output sizes can come from parameters or ports. If they come from ports, the block captures the values when vStart is true in the input control bus. If you select the antialiasing filter, the input stream is filtered with a 5-by-5 fixed kernel before entering the input buffer. The input buffer consists of circular buffers that form the vertical kernels required for the interpolation filter. The interpolator contains vertical and horizontal filters that use the interpolation method that you selected. The antialiasing filter and the interpolation filters use symmetric padding. At the output of the block, the pixelcontrol bus control signals are regenerated, and data is buffered again if necessary to maintain the specified minimum horizontal blanking interval.

Performance

This table shows the resource use after synthesis of the block for the Xilinx^® Zynq^®-7000 SoC ZC706 Evaluation Kit with single-pixel uint8 input and the default parameter settings. The design achieves a clock frequency of 283 MHz.

Resource	Usage
Slice LUTs	1004
Slice Registers	1995
DSP48	10
Block RAM	6

The size of accumulators and buffer memory is determined by the difference in size of the input and output images.

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using Simulink® Coder™.

This block supports C/C++ code generation for Simulink^® accelerator and rapid accelerator modes and for DPI component generation.

HDL Code Generation

Generate VHDL, Verilog and SystemVerilog code for FPGA and ASIC designs using HDL Coder™.

HDL Coder™ provides additional configuration options that affect HDL implementation and synthesized logic.

HDL Architecture

This block has one default HDL architecture.

HDL Block Properties

ConstrainedOutputPipeline	Number of registers to place at the outputs by moving existing delays within your design. Distributed pipelining does not redistribute these registers. The default is 0. For more details, see ConstrainedOutputPipeline (HDL Coder).
InputPipeline	Number of input pipeline stages to insert in the generated code. Distributed pipelining and constrained output pipelining can move these registers. The default is 0. For more details, see InputPipeline (HDL Coder).
OutputPipeline	Number of output pipeline stages to insert in the generated code. Distributed pipelining and constrained output pipelining can move these registers. The default is 0. For more details, see OutputPipeline (HDL Coder).

Restrictions

You cannot generate HDL for this block inside a Resettable Synchronous Subsystem (HDL Coder).

Version History

Introduced in R2023a