pixpark / gpupixel Goto Github PK

Hello, can you share sample android studio project? thank you.

building examples/desktop via CLion on Windows

I can successfully build application(after building the library)

When running with demo image everything is ok, but when I try to load another image instead of your demo.jpg file, I get Segfault at the line gpuSourceImage = SourceImage::create("other_img.jpg").

After some reseach it was found out that segfault occurs in the SourceImage::init function when trying to call glTexImage2D in line 60 when loading image into opengl.

can you tell me what to do to fix this error? not much experience with opengl

thanks

人脸关键点检测使用VNN后，关键点的下标和 Face++是不一样的吧，我看最新代码使用的应该是 104 个关键点模式：
～～～
VNN_Set_Face_Attr(vnn_handle_, "_use_278pts", &use_278pts);
～～～
从 VNN 文档来看104 个关键点分布如下：

face++的 106 个关键点分布如下：

可以看到两种标注的下标是不一样的。

使用 VNN 后在美妆时getFaceIndexs、faceTextureCoordinates并没有同步修改为 VNN 标注的点，LipstickFilter的特效是如何生效的呢？

使用cmake编译总是报错语法错误，请教下如何使用vs导入项目进行编译，多谢

在android studio上起来了，项目也能跑起来，但是起来后就闪退，报错
mac电脑，华为真机测试机，报错如下，请指导一下
FATAL EXCEPTION: main
Process: com.pixpark.gpupixeldemo, PID: 27028
java.lang.RuntimeException: Unable to start activity ComponentInfo{com.pixpark.gpupixeldemo/com.pixpark.GPUPixelApp.MainActivity}: android.view.InflateException: Binary XML file line #14 in com.pixpark.gpupixeldemo:layout/activity_main: Binary XML file line #14 in com.pixpark.gpupixeldemo:layout/activity_main: Error inflating class com.pixpark.gpupixel.GPUPixelView
at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:4765)
at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:4983)
at android.app.servertransaction.LaunchActivityItem.execute(LaunchActivityItem.java:123)
at android.app.servertransaction.TransactionExecutor.executeCallbacks(TransactionExecutor.java:149)
at android.app.servertransaction.TransactionExecutor.execute(TransactionExecutor.java:103)
at android.app.ActivityThread$H.handleMessage(ActivityThread.java:3059)
at android.os.Handler.dispatchMessage(Handler.java:117)
at android.os.Looper.loopOnce(Looper.java:205)
at android.os.Looper.loop(Looper.java:293)
at android.app.ActivityThread.loopProcess(ActivityThread.java:9934)
at android.app.ActivityThread.main(ActivityThread.java:9923)
at java.lang.reflect.Method.invoke(Native Method)
at com.android.internal.os.RuntimeInit$MethodAndArgsCaller.run(RuntimeInit.java:586)
at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:1240)
Caused by: android.view.InflateException: Binary XML file line #14 in com.pixpark.gpupixeldemo:layout/activity_main: Binary XML file line #14 in com.pixpark.gpupixeldemo:layout/activity_main: Error inflating class com.pixpark.gpupixel.GPUPixelView
Caused by: android.view.InflateException: Binary XML file line #14 in com.pixpark.gpupixeldemo:layout/activity_main: Error inflating class com.pixpark.gpupixel.GPUPixelView
Caused by: java.lang.reflect.InvocationTargetException
at java.lang.reflect.Constructor.newInstance0(Native Method)
at java.lang.reflect.Constructor.newInstance(Constructor.java:343)
at android.view.LayoutInflater.createView(LayoutInflater.java:868)
at android.view.LayoutInflater.createViewFromTag(LayoutInflater.java:1020)
at android.view.LayoutInflater.createViewFromTag(LayoutInflater.java:975)
at android.view.LayoutInflater.rInflate(LayoutInflater.java:1137)
at android.view.LayoutInflater.rInflateChildren(LayoutInflater.java:1098)
at android.view.LayoutInflater.inflate(LayoutInflater.java:696)
at android.view.LayoutInflater.inflate(LayoutInflater.java:540)
at com.android.internal.policy.HwPhoneLayoutInflater.inflate(HwPhoneLayoutInflater.java:138)
at com.pixpark.gpupixeldemo.databinding.ActivityMainBinding.inflate(ActivityMainBinding.java:90)
at com.pixpark.gpupixeldemo.databinding.ActivityMainBinding.inflate(ActivityMainBinding.java:84)
at com.pixpark.GPUPixelApp.MainActivity.onCreate(MainActivity.java:48)
at android.app.Activity.performCreate(Activity.java:8592)
at android.app.Activity.performCreate(Activity.java:8565)
at android.app.Instrumentation.callActivityOnCreate(Instrumentation.java:1344)
at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:4733)
at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:4983)
at android.app.servertransaction.LaunchActivityItem.execute(LaunchActivityItem.java:123)
at android.app.servertransaction.TransactionExecutor.executeCallbacks(TransactionExecutor.java:149)
at android.app.servertransaction.TransactionExecutor.execute(TransactionExecutor.java:103)
at android.app.ActivityThread$H.handleMessage(ActivityThread.java:3059)
at android.os.Handler.dispatchMessage(Handler.java:117)
at android.os.Looper.loopOnce(Looper.java:205)
at android.os.Looper.loop(Looper.java:293)
at android.app.ActivityThread.loopProcess(ActivityThread.java:9934)
at android.app.ActivityThread.main(ActivityThread.java:9923)
at java.lang.reflect.Method.invoke(Native Method)
at com.android.internal.os.RuntimeInit$MethodAndArgsCaller.run(RuntimeInit.java:586)
2024-04-30 17:33:15.191 27028-27028 AndroidRuntime com.pixpark.gpupixeldemo E at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:1240)
Caused by: java.lang.UnsatisfiedLinkError: dlopen failed: "/data/app/~~IFugzIY0v7TP8tkOdOCMiw==/com.pixpark.gpupixeldemo-0EPtT2Ej7lL5RGIQylmt0w==/lib/arm64/libgpupixel.so" is 32-bit instead of 64-bit
at java.lang.Runtime.loadLibrary0(Runtime.java:1077)
at java.lang.Runtime.loadLibrary0(Runtime.java:998)
at java.lang.System.loadLibrary(System.java:1657)
at com.pixpark.gpupixel.GPUPixel.(GPUPixel.java:160)
at com.pixpark.gpupixel.GPUPixel.getInstance(GPUPixel.java:48)
at com.pixpark.gpupixel.GPUPixelView.init(GPUPixelView.java:41)
at com.pixpark.gpupixel.GPUPixelView.(GPUPixelView.java:36)
... 30 more

Describe the bug
A clear and concise description of the bug.

How To Reproduce the bug
Steps to reproduce the behavior, how frequent can you experience the bug:
1.

Expected behavior
A clear and concise description of what you expected to happen.

Screenshots
If applicable, add screenshots to help explain your problem.

Running Information(please complete the following information):

• GPUPixel Version using : [e.g. 1.2.1]
• Operating System: [e.g. Ubuntu 20.04, Windows 11, iOS 15.2, Android 13]
• Device Info: [e.g. Huawei Mate40, Xiaomi 10]

Additional context
Add any other context about the problem here.

尝试把gpupixel接入一个安卓项目。首先使用example中的GPUPixelSourceCamera类可以正常工作。想拿到输出图像的数据做进一步处理。之前在windows平台集成sdk使用sourceImage输入处理画面并用captureAProcessedFrameData()方法也可以正常工作。在安卓中调用几帧后崩溃，附调用代码和堆栈图。请问是否可以指点一下

在影视行业，有大量使用python 和 C++ 联合编程的案例来弥补python 在图形图像处理的性能缺陷。这个项目能否在未来有python api 接口，方便我们这些python 开发来将这个模块集成到特效软件中？

An error occurred while compiling under MacOS

path: example/build

error:

/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX14.4.sdk/System/Library/Frameworks/Foundation.framework/Headers/NSObject.h:9:1: error: 
      expected unqualified-id
@class NSInvocation, NSMethodSignature, NSCoder, NSString, NSEnumerator;

想请教一下作者，瘦脸算法是否可以支持实时视频使用，我尝试使用液波或本地平移算法对视频都没有很好的效果，作者有使用过类似的实时瘦脸算法吗，非常感谢

Hi @LiamKeh ,

First of all, GPUPixel is very interesting, I'm trying to integrate GPUPixel into an Open Source WebRTC Video Conferencing application on MacOS, iOS and Android. I'm starting with MacOS first, and a few problems occurred. I have created a RTCBeautyFilter class, it looks like this:

#import "RTCBeautyFilter.h"
#import "gpupixel/gpupixel.h"

using namespace gpupixel;

@interface RTCBeautyFilter () {
    std::shared_ptr<SourceRawDataInput> gpuPixelRawInput;
    GPUPixelView *gpuPixelView;
    std::shared_ptr<BeautyFaceFilter> beauty_face_filter_;
    std::shared_ptr<TargetRawDataOutput> targetRawOutput_;
    std::shared_ptr<FaceReshapeFilter> face_reshape_filter_;
    std::shared_ptr<gpupixel::FaceMakeupFilter> lipstick_filter_;
    std::shared_ptr<gpupixel::FaceMakeupFilter> blusher_filter_;
}

@end

@implementation RTCBeautyFilter

- (instancetype)init {
    self = [super init];
    if (self) {
        [self setup];
    }
    return self;
}

- (void)setup {
    // Init video filter
    [self initVideoFilter];
}

- (void)initVideoFilter {
    gpuPixelRawInput = SourceRawDataInput::create();
 
    // Create filters
    lipstick_filter_ = LipstickFilter::create();
    blusher_filter_ = BlusherFilter::create();
    face_reshape_filter_ = FaceReshapeFilter::create();
    
    gpuPixelRawInput->RegLandmarkCallback([=](std::vector<float> landmarks) {
       lipstick_filter_->SetFaceLandmarks(landmarks);
       blusher_filter_->SetFaceLandmarks(landmarks);
       face_reshape_filter_->SetFaceLandmarks(landmarks);
     });
 
    // Create filter
    targetRawOutput_ = TargetRawDataOutput::create();
    beauty_face_filter_ = BeautyFaceFilter::create();
  
    gpuPixelRawInput->addTarget(lipstick_filter_)
                       ->addTarget(blusher_filter_)
                       ->addTarget(face_reshape_filter_)
                       ->addTarget(beauty_face_filter_)
                       ->addTarget(gpuPixelView);
  
    [gpuPixelView setFillMode:(gpupixel::TargetView::PreserveAspectRatioAndFill)];
}

#pragma mark - Property assignment

- (void)setBeautyValue:(CGFloat)value {
    _beautyValue = value;
    beauty_face_filter_->setBlurAlpha(value/10);
}

- (void)setWhithValue:(CGFloat)value {
    _whithValue = value;
    beauty_face_filter_->setWhite(value/20);
}

- (void)setThinFaceValue:(CGFloat)value {
    _thinFaceValue = value;
    face_reshape_filter_->setFaceSlimLevel(value/100);
}

- (void)setEyeValue:(CGFloat)value {
    _eyeValue = value;
    face_reshape_filter_->setEyeZoomLevel(value/50);
}

- (void)setLipstickValue:(CGFloat)value {
    _lipstickValue = value;
    lipstick_filter_->setBlendLevel(value/10);
}

- (void)setBlusherValue:(CGFloat)value {
    _blusherValue = value;
    blusher_filter_->setBlendLevel(value/10);
}

- (void)processVideoFrame:(CVImageBufferRef)videoFrame {
    // Process video frame here
    CVPixelBufferLockBaseAddress(videoFrame, 0);
    auto width = CVPixelBufferGetWidth(videoFrame);
    auto height = CVPixelBufferGetHeight(videoFrame);
    auto stride = CVPixelBufferGetBytesPerRow(videoFrame)/4;
    auto pixels = (const uint8_t *)CVPixelBufferGetBaseAddress(videoFrame);
    gpuPixelRawInput->uploadBytes(pixels, width, height, stride);
    CVPixelBufferUnlockBaseAddress(videoFrame, 0);
}

@end

and when I execute, the error appears:

ld: warning: Could not find or use auto-linked framework 'CoreAudioTypes': framework 'CoreAudioTypes' not found
ld: Undefined symbols:
  gpupixel::BlusherFilter::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::LipstickFilter::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::BeautyFaceFilter::setBlurAlpha(float), referenced from:
      -[RTCBeautyFilter setBeautyValue:] in RTCBeautyFilter.o
  gpupixel::BeautyFaceFilter::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::BeautyFaceFilter::setWhite(float), referenced from:
      -[RTCBeautyFilter setWhithValue:] in RTCBeautyFilter.o
  gpupixel::FaceMakeupFilter::SetFaceLandmarks(std::__1::vector<float, std::__1::allocator<float>>), referenced from:
      -[RTCBeautyFilter initVideoFilter]::$_0::operator()(std::__1::vector<float, std::__1::allocator<float>>) const in RTCBeautyFilter.o
      -[RTCBeautyFilter initVideoFilter]::$_0::operator()(std::__1::vector<float, std::__1::allocator<float>>) const in RTCBeautyFilter.o
  gpupixel::FaceReshapeFilter::setEyeZoomLevel(float), referenced from:
      -[RTCBeautyFilter setEyeValue:] in RTCBeautyFilter.o
  gpupixel::FaceReshapeFilter::SetFaceLandmarks(std::__1::vector<float, std::__1::allocator<float>>), referenced from:
      -[RTCBeautyFilter initVideoFilter]::$_0::operator()(std::__1::vector<float, std::__1::allocator<float>>) const in RTCBeautyFilter.o
  gpupixel::FaceReshapeFilter::setFaceSlimLevel(float), referenced from:
      -[RTCBeautyFilter setThinFaceValue:] in RTCBeautyFilter.o
  gpupixel::FaceReshapeFilter::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::SourceRawDataInput::uploadBytes(unsigned char const*, int, int, int, long long), referenced from:
      -[RTCBeautyFilter processVideoFrame:] in RTCBeautyFilter.o
  gpupixel::SourceRawDataInput::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::TargetRawDataOutput::create(), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o
  gpupixel::Source::RegLandmarkCallback(std::__1::function<void (std::__1::vector<float, std::__1::allocator<float>>)>), referenced from:
      -[RTCBeautyFilter initVideoFilter] in RTCBeautyFilter.o

Is it correct to integrate GPUPixel into a WebRTC application? and how to fix the above error

你好，我看到示例中有口红和腮红的贴图，但是bounds是写死的，这个依据是什么？

Not Huawei Meta30 . it's Mate30

iOS相机的清晰度不高，与原相机相比，比较模糊

Describe the bug
A clear and concise des

cription of the bug.

运行后内存暴增导致OOM

As titled.

2 issues:

1. Linux demo can't be linked as VNN lib directory is not added in link path. To fix it, please refer attached picture.
2. The demo can run, but the FB doesn't take effect, the output image is not beatified. Please refer the attached picture.
   
   

Platform/Device: Android 14 - Xiaomi 11T 5G

GPUPixel Version: 1.0.6

I use GpuPixel in WebRTC, each VideoFrame processes GpuPixel to apply beauty filter. Below is my code:

fun uploadSourceToGPUPixel(originalBitmap: Bitmap) {
        try {
            if (originalBitmap.width > 0) {
                processingFrame = true

                runBlocking {
                    val sourceImage = GPUPixelSourceImage(originalBitmap)

                    sourceImage.captureAProcessedFrameData(beautyFaceFilter) { result ->
                        Log.d(tag, "---------FRAME PROCESSED---------")
                        callback?.onResult(result ?: originalBitmap)
                        processingFrame = false
                    }
                }
            }
        } catch (error: Exception) {
            Log.e(tag, error.toString())
        }
    }

But my app crash with error logs:

2024-02-15 16:45:43.023  1625-13322 AISPQSelector           camerahalserver                      E  [DpEngineAALReConfigFrame] aispq_en: 0
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  Build fingerprint: 'Xiaomi/amber_global/amber:14/UP1A.231005.007/V816.0.3.0.UKWMIXM:user/release-keys'
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  Revision: '0'
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  ABI: 'arm64'
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  Timestamp: 2024-02-15 16:45:42.917734998+0700
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  Process uptime: 4351s
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  Cmdline: com.waterbus.wanted
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  pid: 9825, tid: 13815, name: Thread-24  >>> com.waterbus.wanted <<<
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  uid: 10729
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0xffffffff00000008
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x0  b4000079977a1180  x1  000000799502c748  x2  0000000000000000  x3  0000000000000000
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x4  0000000000000000  x5  0000000000000000  x6  0000000000000043  x7  0000000000000000
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x8  ffffffff00000000  x9  0000000000000000  x10 ffffffffffffffff  x11 0000000000000000
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x12 0000000000000000  x13 000000799502b2d0  x14 0000007af251ea01  x15 00000000ebad6a89
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x16 00000079745f6eb0  x17 0000007974518acc  x18 0000007947fee000  x19 b4000079ba75b700
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x20 b400007a43c182a0  x21 00000000000001e0  x22 0000000000000168  x23 b4000079ba75b700
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x24 000000799502d000  x25 000000799502c8c0  x26 000000799502c8e4  x27 000000799502c8c0
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      x28 000000799502c7b0  x29 000000799502c730
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A      lr  0000007974542cb4  sp  000000799502c710  pc  0000007974522ec4  pst 0000000040001000
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  14 total frames
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A  backtrace:
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #00 pc 0000000000071ec4  /data/app/~~4K4ZWsbcPj-USsnBCD1-cQ==/com.waterbus.wanted-vsBh_6tS-2BV5IhoR44pWA==/base.apk!libgpupixel.so (offset 0x1c11000) (std::__ndk1::__shared_ptr_pointer<GPUPixel::Filter*, std::__ndk1::default_delete<GPUPixel::Filter>, std::__ndk1::allocator<GPUPixel::Filter> >::__on_zero_shared()+12) (BuildId: 59441ca5dcd7bf4287e6ee971a172bd7897bc944)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #01 pc 0000000000091cb0  /data/app/~~4K4ZWsbcPj-USsnBCD1-cQ==/com.waterbus.wanted-vsBh_6tS-2BV5IhoR44pWA==/base.apk!libgpupixel.so (offset 0x1c11000) (GPUPixel::Source::captureAProcessedFrameData(std::__ndk1::shared_ptr<GPUPixel::Filter>, int, int)+268) (BuildId: 59441ca5dcd7bf4287e6ee971a172bd7897bc944)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #02 pc 00000000000713cc  /data/app/~~4K4ZWsbcPj-USsnBCD1-cQ==/com.waterbus.wanted-vsBh_6tS-2BV5IhoR44pWA==/base.apk!libgpupixel.so (offset 0x1c11000) (Java_com_pixpark_gpupixel_GPUPixel_nativeSourceCaptureAProcessedFrameData+116) (BuildId: 59441ca5dcd7bf4287e6ee971a172bd7897bc944)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #03 pc 0000000000355630  /apex/com.android.art/lib64/libart.so (art_quick_generic_jni_trampoline+144) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #04 pc 00000000005bb68c  /apex/com.android.art/lib64/libart.so (nterp_helper+3852) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #05 pc 000000000133ae8a  /data/app/~~4K4ZWsbcPj-USsnBCD1-cQ==/com.waterbus.wanted-vsBh_6tS-2BV5IhoR44pWA==/oat/arm64/base.vdex (com.pixpark.gpupixel.GPUPixelSource$5.run+130)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #06 pc 00000000005bc4f4  /apex/com.android.art/lib64/libart.so (nterp_helper+7540) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #07 pc 000000000133a4fc  /data/app/~~4K4ZWsbcPj-USsnBCD1-cQ==/com.waterbus.wanted-vsBh_6tS-2BV5IhoR44pWA==/oat/arm64/base.vdex (com.pixpark.gpupixel.GPUPixel$3.run+4)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #08 pc 0000000000505818  /data/misc/apexdata/com.android.art/dalvik-cache/arm64/boot.oat (java.lang.Thread.run+72)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #09 pc 000000000033eba4  /apex/com.android.art/lib64/libart.so (art_quick_invoke_stub+612) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #10 pc 000000000023a9ac  /apex/com.android.art/lib64/libart.so (art::ArtMethod::Invoke(art::Thread*, unsigned int*, unsigned int, art::JValue*, char const*)+144) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #11 pc 000000000053b96c  /apex/com.android.art/lib64/libart.so (art::Thread::CreateCallback(void*)+1600) (BuildId: 02bec5940be704b863f6514fc7d81c41)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #12 pc 0000000000101d5c  /apex/com.android.runtime/lib64/bionic/libc.so (__pthread_start(void*)+204) (BuildId: 84a42637b3a421b801818f5793418fca)
2024-02-15 16:45:43.057 13818-13818 DEBUG                   pid-13818                            A        #13 pc 0000000000095bc0  /apex/com.android.runtime/lib64/bionic/libc.so (__start_thread+64) (BuildId: 84a42637b3a421b801818f5793418fca)

Openface和VNN比怎么样呢？用OpenFace替换掉行吗？

I tried to run on android device ( Pixel 2 ). After granting camera access, the application crashed, displaying the following errors.

I try to run it in Android Studio with emulator, but it runs failed. The root cause is: in 'build.gradle', it defines targetSdk to '31', but in the code (GPUPixelSourceCamera.java) is still using the 'Camera' APIs which have been deprecated. Or anything am I doing wrong?

public final class BeautyFaceFilter {
public static final String name = "BeautyFaceFilter";
public static final String propSmoothLevel = "skin_smoothing";
public static final String propWhiteLevel = "whiteness";
}
设置的属性里只有这两种，没有其他的。

在源码中并未发现glfw3.lib，尝试下载3.3.8版本库并在VS2022中设置了引用，报错：

是否需要别的操作解决？

目前实现了瘦脸、大小眼、口红、晒红，有眉毛、牙齿的美颜滤镜处理吗

repo 的实时视频和图像的能力令人印象深刻，在研究您的项目时，我对其中的 FaceMakeupFilter 功能产生了一些疑问，希望能得到您的解答。

在 DEMO 中有实现两个 Filter，分别是blusher_filter_和lipstick_filter_，它们是如何与landmarks协同去定位面部位置的呢？这里有几点没有理解：

• 从代码看它们使用的是同一个类FaceMakeupFilter，唯一的差别是setTextureBounds的位置不同，是不是意味着 filters 并不知道什么叫嘴什么叫腮，它仅仅是根据TextureBounds去做形变？
• 这两个滤镜的TextureBounds FrameBounds{502.5, 710, 262.5, 167.5}和FrameBounds{395, 520, 489, 209}是如何得到的呢？如果我想要去实现一个新的滤镜，该如何确定这些值？
• 我试了一下口红滤镜，它能够准确避免牙齿被上色，这是似乎是通过faceIndexs对嘴部的三角形进行着色实现的？但是mouth 是一张图片，并不是一个颜色，它是如何形变去适配嘴部区域的？尤其是考虑到前面的第一点：FaceMakeupFilter应该并不知道它的实例的想要美化的是嘴还是腮？

I'm thoroughly impressed with your real-time video and image processing library on GitHub. While exploring your project, I have some questions about the FaceMakeupFilter feature and would appreciate your insights.

In the DEMO, two filters are implemented: blusher_filter_ and lipstick_filter_. I'm intrigued by how they work in conjunction with facial landmarks. Here are my questions:

• Both filters seem to use the same FaceMakeupFilter class, with the main difference being the placement of setTextureBounds. Does this mean that the filters don't directly recognize specific facial areas like the mouth or cheeks, and instead deform based on TextureBounds?
• How are the TextureBounds for blusher_filter_ and lipstick_filter_ determined? For instance, how were values like FrameBounds{502.5, 710, 262.5, 167.5} and FrameBounds{395, 520, 489, 209} arrived at? If I want to implement a new filter, how should I determine the values for these TextureBounds?
• I tried the lipstick filter and noticed its precise avoidance of coloring the teeth, apparently achieved by coloring the mouth's triangles using faceIndexs. However, if the mouth is represented as an image rather than a single color, how does it deform to fit the mouth area? This is particularly intriguing considering FaceMakeupFilter might not differentiate whether it's being used for the mouth or cheeks.

对图片的处理用前台GLSurfaceView来渲染,不太方便,是否能增加离屏渲染处理的Demo.

pid: 5217, tid: 5250, name: GLThread 325 >>> com.pixpark.gpupixel <<<
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x108
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x0 0000000000000000 x1 0000000000000006 x2 0000000000000000 x3 0000000000000006
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x4 0000007554ffde48 x5 000000754146052a x6 0000000000000000 x7 0000000000000000
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x8 0000000000000000 x9 0000000000000006 x10 0000000000430000 x11 0000007554fff4e8
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x12 00000000000000c4 x13 0000000000000012 x14 000000756083a250 x15 dfcb56e8fb5eb3f5
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x16 0000007540f69638 x17 0000007540eb81b4 x18 000000006f56dc08 x19 000000755c29f000
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x20 000000755bca9e50 x21 000000755c29f000 x22 0000007554ffdc4c x23 000000754146052a
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x24 000000000000000c x25 17e56458624f2ffe x26 000000755c29f098 x27 17e56458624f2ffe
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A x28 0000000000000003 x29 0000007554ffd980 x30 0000007540e2b1d4
2023-12-07 14:12:31.612 5334-5334 DEBUG pid-5334 A sp 0000007554ffd950 pc 0000007540eb81c8 pstate 0000000020000000
2023-12-07 14:12:31.614 5334-5334 DEBUG pid-5334 A
backtrace:
2023-12-07 14:12:31.614 5334-5334 DEBUG pid-5334 A #00 pc 00000000001311c8 /data/app/com.pixpark.gpupixel-1/lib/arm64/libgpupixel.so (_ZN8GPUPixel18SourceRawDataInput11setRotationENS_12RotationModeE+20)
2023-12-07 14:12:31.614 5334-5334 DEBUG pid-5334 A #1 pc 00000000000a41d0 /data/app/com.pixpark.gpupixel-1/lib/arm64/libgpupixel.so (Java_com_pixpark_gpupixel_GPUPixel_nativeSourceRawInputSetRotation+40)
2023-12-07 14:12:31.614 5334-5334 DEBUG pid-5334 A #2 pc 000000000069cccc /data/app/com.pixpark.gpupixel-1/oat/arm64/base.odex (offset 0x603000)

beauty_face_filter设置white value 会花屏

如题，担心以后恐怕得iOS安卓维护两套

作者你好，我想咨询下如果我的需求是只使用图片进行这些滤镜的处理而不用取景框，并且能保存处理后的图片到本地，这个功能在现有的demo里可以实现吗
我有发现有一个GPUPixelSourceImage是可以直接传入图片的bitmap的，但是GPUPixelView这个类中只添加了surfaceview，是否我需要参照这个添加一个类只有一个imageview的子布局呢

如题
现在项目中的拍摄功能，使用 GPUImage 进行滤镜渲染之类的。
是否可以同时 使用 gpupixel进行 嘴唇 大眼等效果。
麻烦了。

如題
希望能指點一下

这个要怎么输出成mov或者mp4等格式的视频

wsl2 下ubuntu20.04

iOS 使用 captureAProcessedFrameData这个函数返回的是个什么结果呢？如何转换成图片
安卓调用这个api，目前会发生崩溃

Hi,
请问工程中有glsl文件吗？read.me中我看了工程调用了opengl的相关库。

Can you build a sample using GPUPixelSourceRawInput, TargetRawDataOutput? I see the iOS sample is quite comprehensive, but for Android, there's only a sample using the camera.
In this sample i have to use GPUPixelView to make sure that renderer can call onDraw, when i try to replace with TargetRawDataOutput. onDraw can't invoke

我想添加新的face_makeup_filter，增加处理牙齿、眉毛、美瞳这些效果，请问是怎么创建这些素材，是要有一个标准的人脸图在上面设计的吗？

Is your feature request related to a problem? Please describe.
无法在arm架构下的linux系统进行编译

Describe the solution you'd like
Copying headers and resource to output directory.
/usr/bin/ld: ../../src/third_party/vnn/libs/linux/libvnn_core.so: error adding symbols: file in wrong format
collect2: error: ld returned 1 exit status
make[2]: *** [src/CMakeFiles/gpupixel.dir/build.make:1832：../output/library/linux/libgpupixel.so] 错误 1
make[1]: *** [CMakeFiles/Makefile2:142：src/CMakeFiles/gpupixel.dir/all] 错误 2
make: *** [Makefile:84：all] 错误 2

Describe alternatives you've considered
在third_party/vnn/libs/linux的库是x86架构的，是否增加arm架构的，或者放一下源码路径，自己编译

Additional context
Add any other context or screenshots about the feature request here.

测试了一张多个人脸的图片，从左到右只有第一个人脸生效，不支持多个人脸吗？

在 Android Demo 工程中，Pad 横屏下预览未进行旋转，查看代码 com.pixpark.gpupixel.GPUPixelSourceCamera#mRotation 赋值后并未使用，辛苦确认一下如何使用。

Face++ has limitations and many features require the use of landmarks.