WesleyZhang1991 / acf

Aggregated Channel Feature object detection in C++ and OpenGL ES 2.0

acf

Aggregated Channel Feature object detection in C++ and OpenGL ES 2.0

This module is very well suited to running real time object detection on mobile processors, where recent high performing but GPU needy DNN approaches aren't as suitable. The ACF pyramids can be computed with the OpenGL ES 2.0 shaders and retrieved more or less for free (< 1 frame time with 1 frame of latency). For selfie video, the pretrained face detectors (see FACE80 and FACE64) run in a few milliseconds on an iPhone 7. TODO: The Locally Decorrelated Channel Feature addition has not yet been added (see LDCF), but the 5x5 kernels should map well to OpenGL shaders. That should make performance very competitive (see Piotr's references for comparisons).

Sample 10 Channel ACF from GPU: LUV + magnitude (locally normalized) + gradient orientation (6 bins):

• C++ implementation of Fast Feature Pyramids for Object Detection (see Piotr's Matlab Toolbox)
• Hunter package management for cross platform builds: "Organized Freedom!" :)
• Fast OpenGL ES 2.0 ACF Pyramid computation via ogles_gpgpu
• OpenCV API

Quick Start

ACF is a CMake based project that uses the Hunter package manager to download and build project dependencies from source as needed. Hunter contains detailed documentation, but a few high level notes and documentation links are provided here to help orient first time users. In practice, some working knowledge of CMake may also be required. Hunter itself is written in CMake, and is installed as part of the build process from a single HunterGate() macro at the top of the root CMakeLists.txt file (typically cmake/Hunter/HunterGate.cmake) (you don't have to build or install it). Each CMake dependency's find_package(FOO) call that is paired with a hunter_add_package(FOO CONFIG REQUIRED) will be managed by Hunter. In most cases, the only system requirement for building a Hunter project is a recent CMake with CURL support and a working compiler correpsonding to the operative toolchain. Hunter will maintain all dependencies in a versioned local cache by default (typically ${HOME}/.hunter) where they can be reused in subsequent builds and shared between different projects. They can also be stored in a server side binary cache -- select toolchains will be backed by a server side binary cache (https://github.com/elucideye/hunter-cache) and will produce faster first time builds (use them if you can!).

The Travis (Linux/OSX/iOS/Android) and Appveyor (Windows) CI scripts in the project's root directory can serve as a reference for basic setup when building from source. To support cross platform builds and testing, the CI scripts make use of Polly: a set of common CMake toolchains paired with a simple polly.py CMake build script. Polly is used here for convenience to generate CMake command line invocations -- it is not required for building Hunter projects.

To reproduce the CI builds on a local host, the following setup is recommended:

• Install compiler: http://cgold.readthedocs.io/en/latest/first-step.html
• Install CMake (and add to PATH)
• Install Python (for Polly)
• Clone Polly and add <polly>/bin to PATH

Note: Polly is not a build requirement, CMake can always be used directly, but it is used here for convenience.

The bin/hunter_env.{sh,cmd} scripts (used in the CI builds) can be used as a fast shortcut to install these tools for you. You may want to add the PATH variables permanently to your .bashrc file (or equivalent) for future sessions.

Linux/OSX/Android/iOS	Windows
source bin/hunter_env.sh	bin\hunter_env.cmd

After the environment is configured, you can build for any supported Polly toolchain (see polly.py --help) with a command like this:

polly.py --toolchain ${TOOLCHAIN} --config ${CONFIG} --fwd HUNTER_CONFIGURATION_TYPES=${CONFIG} --install --verbose

HOWTO

_install/
└── libcxx
    ├── bin
    │   ├── acf-detect
    │   └── acf-mat2cpb
    ├── include
    │   └── acf
    │       ├── ACF.h
    │       ├── ACFField.h
    │       ├── GPUACF.h
    │       ├── MatP.h
    │       ├── ObjectDetector.h
    │       ├── acf_common.h
    │       └── acf_export.h
    └── lib
         ├── cmake
         │   └── acf
         │       ├── acfConfig.cmake
         │       ├── acfConfigVersion.cmake
         │       ├── acfTargets-release.cmake
         │       └── acfTargets.cmake
         └── libacf.a

$ cd _install/${TOOLCHAIN}/bin
$ wget https://github.com/elucideye/drishti-assets/raw/master/drishti_face_gray_80x80.cpb
$ wget https://github.com/elucideye/drishti-faces/raw/master/lena512color.png
$ ./acf-detect --input=lena512color.png --output=/tmp/ --model=drishti_face_gray_80x80.cpb --nms --annotate --calibration=0.00001

[16:56:34.092 | thread:8703967691101883897 | acf-detect | info]: 1/1 /Users/dhirvonen/devel/elucideye//drishti-faces/lena512color.png = 1; score = 26.0038