Home

Awesome

GPUImage

<div style="float: right"><img src="http://sunsetlakesoftware.com/sites/default/files/GPUImageLogo.png" /></div>

<a href="https://zenodo.org/record/10416#.U5YGaF773Md"><img src="https://zenodo.org/badge/doi/10.5281/zenodo.10416.svg" /></a>

Brad Larson

http://www.sunsetlakesoftware.com

@bradlarson

contact@sunsetlakesoftware.com

Overview

The GPUImage framework is a BSD-licensed iOS library that lets you apply GPU-accelerated filters and other effects to images, live camera video, and movies. In comparison to Core Image (part of iOS 5.0), GPUImage allows you to write your own custom filters, supports deployment to iOS 4.0, and has a simpler interface. However, it currently lacks some of the more advanced features of Core Image, such as facial detection.

For massively parallel operations like processing images or live video frames, GPUs have some significant performance advantages over CPUs. On an iPhone 4, a simple image filter can be over 100 times faster to perform on the GPU than an equivalent CPU-based filter.

However, running custom filters on the GPU requires a lot of code to set up and maintain an OpenGL ES 2.0 rendering target for these filters. I created a sample project to do this:

http://www.sunsetlakesoftware.com/2010/10/22/gpu-accelerated-video-processing-mac-and-ios

and found that there was a lot of boilerplate code I had to write in its creation. Therefore, I put together this framework that encapsulates a lot of the common tasks you'll encounter when processing images and video and made it so that you don't need to care about the OpenGL ES 2.0 underpinnings.

This framework compares favorably to Core Image when handling video, taking only 2.5 ms on an iPhone 4 to upload a frame from the camera, apply a gamma filter, and display, versus 106 ms for the same operation using Core Image. CPU-based processing takes 460 ms, making GPUImage 40X faster than Core Image for this operation on this hardware, and 184X faster than CPU-bound processing. On an iPhone 4S, GPUImage is only 4X faster than Core Image for this case, and 102X faster than CPU-bound processing. However, for more complex operations like Gaussian blurs at larger radii, Core Image currently outpaces GPUImage.

License

BSD-style, with the full license available with the framework in License.txt.

Technical requirements

General architecture

GPUImage uses OpenGL ES 2.0 shaders to perform image and video manipulation much faster than could be done in CPU-bound routines. However, it hides the complexity of interacting with the OpenGL ES API in a simplified Objective-C interface. This interface lets you define input sources for images and video, attach filters in a chain, and send the resulting processed image or video to the screen, to a UIImage, or to a movie on disk.

Images or frames of video are uploaded from source objects, which are subclasses of GPUImageOutput. These include GPUImageVideoCamera (for live video from an iOS camera), GPUImageStillCamera (for taking photos with the camera), GPUImagePicture (for still images), and GPUImageMovie (for movies). Source objects upload still image frames to OpenGL ES as textures, then hand those textures off to the next objects in the processing chain.

Filters and other subsequent elements in the chain conform to the GPUImageInput protocol, which lets them take in the supplied or processed texture from the previous link in the chain and do something with it. Objects one step further down the chain are considered targets, and processing can be branched by adding multiple targets to a single output or filter.

For example, an application that takes in live video from the camera, converts that video to a sepia tone, then displays the video onscreen would set up a chain looking something like the following:

GPUImageVideoCamera -> GPUImageSepiaFilter -> GPUImageView

Adding the static library to your iOS project

Note: if you want to use this in a Swift project, you need to use the steps in the "Adding this as a framework" section instead of the following. Swift needs modules for third-party code.

Once you have the latest source code for the framework, it's fairly straightforward to add it to your application. Start by dragging the GPUImage.xcodeproj file into your application's Xcode project to embed the framework in your project. Next, go to your application's target and add GPUImage as a Target Dependency. Finally, you'll want to drag the libGPUImage.a library from the GPUImage framework's Products folder to the Link Binary With Libraries build phase in your application's target.

GPUImage needs a few other frameworks to be linked into your application, so you'll need to add the following as linked libraries in your application target:

You'll also need to find the framework headers, so within your project's build settings set the Header Search Paths to the relative path from your application to the framework/ subdirectory within the GPUImage source directory. Make this header search path recursive.

To use the GPUImage classes within your application, simply include the core framework header using the following:

#import "GPUImage.h"

As a note: if you run into the error "Unknown class GPUImageView in Interface Builder" or the like when trying to build an interface with Interface Builder, you may need to add -ObjC to your Other Linker Flags in your project's build settings.

Also, if you need to deploy this to iOS 4.x, it appears that the current version of Xcode (4.3) requires that you weak-link the Core Video framework in your final application or you see crashes with the message "Symbol not found: _CVOpenGLESTextureCacheCreate" when you create an archive for upload to the App Store or for ad hoc distribution. To do this, go to your project's Build Phases tab, expand the Link Binary With Libraries group, and find CoreVideo.framework in the list. Change the setting for it in the far right of the list from Required to Optional.

Additionally, this is an ARC-enabled framework, so if you want to use this within a manual reference counted application targeting iOS 4.x, you'll need to add -fobjc-arc to your Other Linker Flags as well.

Building a static library at the command line

If you don't want to include the project as a dependency in your application's Xcode project, you can build a universal static library for the iOS Simulator or device. To do this, run build.sh at the command line. The resulting library and header files will be located at build/Release-iphone. You may also change the version of the iOS SDK by changing the IOSSDK_VER variable in build.sh (all available versions can be found using xcodebuild -showsdks).

Adding this as a framework (module) to your Mac or iOS project

Xcode 6 and iOS 8 support the use of full frameworks, as does the Mac, which simplifies the process of adding this to your application. To add this to your application, I recommend dragging the .xcodeproj project file into your application's project (as you would in the static library target).

For your application, go to its target build settings and choose the Build Phases tab. Under the Target Dependencies grouping, add GPUImageFramework on iOS (not GPUImage, which builds the static library) or GPUImage on the Mac. Under the Link Binary With Libraries section, add GPUImage.framework.

This should cause GPUImage to build as a framework. Under Xcode 6, this will also build as a module, which will allow you to use this in Swift projects. When set up as above, you should just need to use

import GPUImage

to pull it in.

You then need to add a new Copy Files build phase, set the Destination to Frameworks, and add the GPUImage.framework build product to that. This will allow the framework to be bundled with your application (otherwise, you'll see cryptic "dyld: Library not loaded: @rpath/GPUImage.framework/GPUImage" errors on execution).

Documentation

Documentation is generated from header comments using appledoc. To build the documentation, switch to the "Documentation" scheme in Xcode. You should ensure that "APPLEDOC_PATH" (a User-Defined build setting) points to an appledoc binary, available on <a href="https://github.com/tomaz/appledoc">Github</a> or through <a href="https://github.com/Homebrew/homebrew">Homebrew</a>. It will also build and install a .docset file, which you can view with your favorite documentation tool.

Performing common tasks

Filtering live video

To filter live video from an iOS device's camera, you can use code like the following:

GPUImageVideoCamera *videoCamera = [[GPUImageVideoCamera alloc] initWithSessionPreset:AVCaptureSessionPreset640x480 cameraPosition:AVCaptureDevicePositionBack];
videoCamera.outputImageOrientation = UIInterfaceOrientationPortrait;

GPUImageFilter *customFilter = [[GPUImageFilter alloc] initWithFragmentShaderFromFile:@"CustomShader"];
GPUImageView *filteredVideoView = [[GPUImageView alloc] initWithFrame:CGRectMake(0.0, 0.0, viewWidth, viewHeight)];

// Add the view somewhere so it's visible

[videoCamera addTarget:customFilter];
[customFilter addTarget:filteredVideoView];

[videoCamera startCameraCapture];

This sets up a video source coming from the iOS device's back-facing camera, using a preset that tries to capture at 640x480. This video is captured with the interface being in portrait mode, where the landscape-left-mounted camera needs to have its video frames rotated before display. A custom filter, using code from the file CustomShader.fsh, is then set as the target for the video frames from the camera. These filtered video frames are finally displayed onscreen with the help of a UIView subclass that can present the filtered OpenGL ES texture that results from this pipeline.

The fill mode of the GPUImageView can be altered by setting its fillMode property, so that if the aspect ratio of the source video is different from that of the view, the video will either be stretched, centered with black bars, or zoomed to fill.

For blending filters and others that take in more than one image, you can create multiple outputs and add a single filter as a target for both of these outputs. The order with which the outputs are added as targets will affect the order in which the input images are blended or otherwise processed.

Also, if you wish to enable microphone audio capture for recording to a movie, you'll need to set the audioEncodingTarget of the camera to be your movie writer, like for the following:

videoCamera.audioEncodingTarget = movieWriter;

Capturing and filtering a still photo

To capture and filter still photos, you can use a process similar to the one for filtering video. Instead of a GPUImageVideoCamera, you use a GPUImageStillCamera:

stillCamera = [[GPUImageStillCamera alloc] init];
stillCamera.outputImageOrientation = UIInterfaceOrientationPortrait;

filter = [[GPUImageGammaFilter alloc] init];
[stillCamera addTarget:filter];
GPUImageView *filterView = (GPUImageView *)self.view;
[filter addTarget:filterView];

[stillCamera startCameraCapture];

This will give you a live, filtered feed of the still camera's preview video. Note that this preview video is only provided on iOS 4.3 and higher, so you may need to set that as your deployment target if you wish to have this functionality.

Once you want to capture a photo, you use a callback block like the following:

[stillCamera capturePhotoProcessedUpToFilter:filter withCompletionHandler:^(UIImage *processedImage, NSError *error){
    NSData *dataForJPEGFile = UIImageJPEGRepresentation(processedImage, 0.8);

    NSArray *paths = NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES);
    NSString *documentsDirectory = [paths objectAtIndex:0];

    NSError *error2 = nil;
    if (![dataForJPEGFile writeToFile:[documentsDirectory stringByAppendingPathComponent:@"FilteredPhoto.jpg"] options:NSAtomicWrite error:&error2])
    {
        return;
    }
}];

The above code captures a full-size photo processed by the same filter chain used in the preview view and saves that photo to disk as a JPEG in the application's documents directory.

Note that the framework currently can't handle images larger than 2048 pixels wide or high on older devices (those before the iPhone 4S, iPad 2, or Retina iPad) due to texture size limitations. This means that the iPhone 4, whose camera outputs still photos larger than this, won't be able to capture photos like this. A tiling mechanism is being implemented to work around this. All other devices should be able to capture and filter photos using this method.

Processing a still image

There are a couple of ways to process a still image and create a result. The first way you can do this is by creating a still image source object and manually creating a filter chain:

UIImage *inputImage = [UIImage imageNamed:@"Lambeau.jpg"];

GPUImagePicture *stillImageSource = [[GPUImagePicture alloc] initWithImage:inputImage];
GPUImageSepiaFilter *stillImageFilter = [[GPUImageSepiaFilter alloc] init];

[stillImageSource addTarget:stillImageFilter];
[stillImageFilter useNextFrameForImageCapture];
[stillImageSource processImage];

UIImage *currentFilteredVideoFrame = [stillImageFilter imageFromCurrentFramebuffer];

Note that for a manual capture of an image from a filter, you need to set -useNextFrameForImageCapture in order to tell the filter that you'll be needing to capture from it later. By default, GPUImage reuses framebuffers within filters to conserve memory, so if you need to hold on to a filter's framebuffer for manual image capture, you need to let it know ahead of time.

For single filters that you wish to apply to an image, you can simply do the following:

GPUImageSepiaFilter *stillImageFilter2 = [[GPUImageSepiaFilter alloc] init];
UIImage *quickFilteredImage = [stillImageFilter2 imageByFilteringImage:inputImage];

Writing a custom filter

One significant advantage of this framework over Core Image on iOS (as of iOS 5.0) is the ability to write your own custom image and video processing filters. These filters are supplied as OpenGL ES 2.0 fragment shaders, written in the C-like OpenGL Shading Language.

A custom filter is initialized with code like

GPUImageFilter *customFilter = [[GPUImageFilter alloc] initWithFragmentShaderFromFile:@"CustomShader"];

where the extension used for the fragment shader is .fsh. Additionally, you can use the -initWithFragmentShaderFromString: initializer to provide the fragment shader as a string, if you would not like to ship your fragment shaders in your application bundle.

Fragment shaders perform their calculations for each pixel to be rendered at that filter stage. They do this using the OpenGL Shading Language (GLSL), a C-like language with additions specific to 2-D and 3-D graphics. An example of a fragment shader is the following sepia-tone filter:

varying highp vec2 textureCoordinate;

uniform sampler2D inputImageTexture;

void main()
{
    lowp vec4 textureColor = texture2D(inputImageTexture, textureCoordinate);
    lowp vec4 outputColor;
    outputColor.r = (textureColor.r * 0.393) + (textureColor.g * 0.769) + (textureColor.b * 0.189);
    outputColor.g = (textureColor.r * 0.349) + (textureColor.g * 0.686) + (textureColor.b * 0.168);    
    outputColor.b = (textureColor.r * 0.272) + (textureColor.g * 0.534) + (textureColor.b * 0.131);
	outputColor.a = 1.0;

	gl_FragColor = outputColor;
}

For an image filter to be usable within the GPUImage framework, the first two lines that take in the textureCoordinate varying (for the current coordinate within the texture, normalized to 1.0) and the inputImageTexture uniform (for the actual input image frame texture) are required.

The remainder of the shader grabs the color of the pixel at this location in the passed-in texture, manipulates it in such a way as to produce a sepia tone, and writes that pixel color out to be used in the next stage of the processing pipeline.

One thing to note when adding fragment shaders to your Xcode project is that Xcode thinks they are source code files. To work around this, you'll need to manually move your shader from the Compile Sources build phase to the Copy Bundle Resources one in order to get the shader to be included in your application bundle.

Filtering and re-encoding a movie

Movies can be loaded into the framework via the GPUImageMovie class, filtered, and then written out using a GPUImageMovieWriter. GPUImageMovieWriter is also fast enough to record video in realtime from an iPhone 4's camera at 640x480, so a direct filtered video source can be fed into it. Currently, GPUImageMovieWriter is fast enough to record live 720p video at up to 20 FPS on the iPhone 4, and both 720p and 1080p video at 30 FPS on the iPhone 4S (as well as on the new iPad).

The following is an example of how you would load a sample movie, pass it through a pixellation filter, then record the result to disk as a 480 x 640 h.264 movie:

movieFile = [[GPUImageMovie alloc] initWithURL:sampleURL];
pixellateFilter = [[GPUImagePixellateFilter alloc] init];

[movieFile addTarget:pixellateFilter];

NSString *pathToMovie = [NSHomeDirectory() stringByAppendingPathComponent:@"Documents/Movie.m4v"];
unlink([pathToMovie UTF8String]);
NSURL *movieURL = [NSURL fileURLWithPath:pathToMovie];

movieWriter = [[GPUImageMovieWriter alloc] initWithMovieURL:movieURL size:CGSizeMake(480.0, 640.0)];
[pixellateFilter addTarget:movieWriter];

movieWriter.shouldPassthroughAudio = YES;
movieFile.audioEncodingTarget = movieWriter;
[movieFile enableSynchronizedEncodingUsingMovieWriter:movieWriter];

[movieWriter startRecording];
[movieFile startProcessing];

Once recording is finished, you need to remove the movie recorder from the filter chain and close off the recording using code like the following:

[pixellateFilter removeTarget:movieWriter];
[movieWriter finishRecording];

A movie won't be usable until it has been finished off, so if this is interrupted before this point, the recording will be lost.

Interacting with OpenGL ES

GPUImage can both export and import textures from OpenGL ES through the use of its GPUImageTextureOutput and GPUImageTextureInput classes, respectively. This lets you record a movie from an OpenGL ES scene that is rendered to a framebuffer object with a bound texture, or filter video or images and then feed them into OpenGL ES as a texture to be displayed in the scene.

The one caution with this approach is that the textures used in these processes must be shared between GPUImage's OpenGL ES context and any other context via a share group or something similar.

Built-in filters

There are currently 125 built-in filters, divided into the following categories:

Color adjustments

Image processing

Blending modes

Visual effects

You can also easily write your own custom filters using the C-like OpenGL Shading Language, as described above.

Sample applications

Several sample applications are bundled with the framework source. Most are compatible with both iPhone and iPad-class devices. They attempt to show off various aspects of the framework and should be used as the best examples of the API while the framework is under development. These include:

SimpleImageFilter

A bundled JPEG image is loaded into the application at launch, a filter is applied to it, and the result rendered to the screen. Additionally, this sample shows two ways of taking in an image, filtering it, and saving it to disk.

SimpleVideoFilter

A pixellate filter is applied to a live video stream, with a UISlider control that lets you adjust the pixel size on the live video.

SimpleVideoFileFilter

A movie file is loaded from disk, an unsharp mask filter is applied to it, and the filtered result is re-encoded as another movie.

MultiViewFilterExample

From a single camera feed, four views are populated with realtime filters applied to camera. One is just the straight camera video, one is a preprogrammed sepia tone, and two are custom filters based on shader programs.

FilterShowcase

This demonstrates every filter supplied with GPUImage.

BenchmarkSuite

This is used to test the performance of the overall framework by testing it against CPU-bound routines and Core Image. Benchmarks involving still images and video are run against all three, with results displayed in-application.

CubeExample

This demonstrates the ability of GPUImage to interact with OpenGL ES rendering. Frames are captured from the camera, a sepia filter applied to them, and then they are fed into a texture to be applied to the face of a cube you can rotate with your finger. This cube in turn is rendered to a texture-backed framebuffer object, and that texture is fed back into GPUImage to have a pixellation filter applied to it before rendering to screen.

In other words, the path of this application is camera -> sepia tone filter -> cube -> pixellation filter -> display.

ColorObjectTracking

A version of my ColorTracking example from http://www.sunsetlakesoftware.com/2010/10/22/gpu-accelerated-video-processing-mac-and-ios ported across to use GPUImage, this application uses color in a scene to track objects from a live camera feed. The four views you can switch between include the raw camera feed, the camera feed with pixels matching the color threshold in white, the processed video where positions are encoded as colors within the pixels passing the threshold test, and finally the live video feed with a dot that tracks the selected color. Tapping the screen changes the color to track to match the color of the pixels under your finger. Tapping and dragging on the screen makes the color threshold more or less forgiving. This is most obvious on the second, color thresholding view.

Currently, all processing for the color averaging in the last step is done on the CPU, so this is part is extremely slow.