COMPILE_TIME_ASSERTs都无法将dlib作为共享库引用

Posted

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了COMPILE_TIME_ASSERTs都无法将dlib作为共享库引用相关的知识,希望对你有一定的参考价值。

我正在尝试为dlib编写一个JNI包装器,以便我可以在Java中获取面部描述符。我已经写了jni代码来做它,但我似乎无法make它(在MacOS X上)。我的JNI不会产生任何编译错误,但是当make-it时,大量的COMPILE_TIME_ASSERTS会失败。

我做错了什么,我怎样才能成功实现这一目标?

的CMakeLists.txt:

cmake_minimum_required(VERSION 2.8.12)

project(dlib-jni)

set(CMAKE_CXX_STANDARD 11)
# add_subdirectory(../dlib dlib_build)
include_directories(src)

find_package(dlib REQUIRED) #possibly supporting components!
# include_directories(${OpenCV_INCLUDE_DIRS})
message(STATUS "Using dlib-${dlib_VERSION}")

find_package(OpenCV 3 REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})

find_package(JNI REQUIRED)
message (STATUS "JNI_INCLUDE_DIRS=${JNI_INCLUDE_DIRS}")
message (STATUS "JAVA_INCLUDE_PATH =${JAVA_INCLUDE_PATH}")
message (STATUS "JNI_LIBRARIES=${JNI_LIBRARIES}")
message (STATUS "JAVA_JVM_LIBRARY=${JAVA_JVM_LIBRARY}")
include_directories(${JNI_INCLUDE_DIRS})

add_library(dlib_jni SHARED src/dlib-jni.cpp)
target_link_libraries(dlib_jni dlib::dlib opencv_core opencv_highgui)

cmake命令:

cmake -DOpenCV_DIR=/usr/local/opt/opencv/share/OpenCV/OpenCVConfig.cmake  ../

cmake输出:

-- The C compiler identification is AppleClang 9.1.0.9020039
-- The CXX compiler identification is AppleClang 9.1.0.9020039
-- Check for working C compiler: /Library/Developer/CommandLineTools/usr/bin/cc
-- Check for working C compiler: /Library/Developer/CommandLineTools/usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /Library/Developer/CommandLineTools/usr/bin/c++
-- Check for working CXX compiler: /Library/Developer/CommandLineTools/usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Using dlib-19.15.99
-- Found OpenCV: /usr/local (found suitable version "3.3.1", minimum required is "3") 
-- Found JNI: /Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/jre/lib/libjawt.dylib  
-- JNI_INCLUDE_DIRS=/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/include;/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/include/darwin;/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/include
-- JAVA_INCLUDE_PATH =/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/include
-- JNI_LIBRARIES=/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/jre/lib/libjawt.dylib;/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/jre/lib/server/libjvm.dylib
-- JAVA_JVM_LIBRARY=/Library/Java/JavaVirtualMachines/jdk1.8.0_131.jdk/Contents/Home/jre/lib/server/libjvm.dylib
-- Configuring done
CMake Warning (dev):
  Policy CMP0042 is not set: MACOSX_RPATH is enabled by default.  Run "cmake
  --help-policy CMP0042" for policy details.  Use the cmake_policy command to
  set the policy and suppress this warning.

  MACOSX_RPATH is not specified for the following targets:

   dlib_jni

This warning is for project developers.  Use -Wno-dev to suppress it.

-- Generating done
-- Build files have been written to: /Users/me/gitrepos/dlib-jni/build

make命令:

make && make install

第一个错误(有很多):

/usr/local/include/dlib/array2d/../geometry/../image_processing/../image_transforms/image_pyramid.h:940:13: error: static_assert failed "Failed assertion"
            COMPILE_TIME_ASSERT( pixel_traits<in_pixel_type>::has_alpha == false );
答案

COMPILE_TIME_ASSERTs用于确保使用dlib的代码实际上是正确的。

碰巧,我的不是。

在我的例子中,我从OpenCV Mat创建了一个dlib cv_image,如下所示:

dlib::cv_image<rgb_pixel_apha> img(image);

然后尝试检测这样的面孔:

std::vector<dlib::matrix<rgb_pixel>> faces; 
for (auto face : detector(img)) 
...

敏锐的开发人员会注意到他们的模板不同。一个是rgb_pixel_alpha,另一个是rgb_pixel

COMPILE_TIME_ASSERT检查alpha通道。我只是没有正确理解错误。

以上是关于COMPILE_TIME_ASSERTs都无法将dlib作为共享库引用的主要内容,如果未能解决你的问题,请参考以下文章

无法将属性绑定到“FlywayProperties”,AnnotationConfigApplicationContext@5454d35e 尚未刷新

AddTransientAddSingletonAddScopped 三者都应该在什么场景下使用

Python numpy:无法将 datetime64[ns] 转换为 datetime64[D](与 Numba 一起使用)

无法将 Python 算法翻译成 C++

无法将 Spring 应用程序连接到 debezium kafka

无法将 unicode .csv 读入 R