👀 openyolo - Object detection and classification

This repository is a fork of the primary repository with minor modifications to run on Raspian Buster and Ubuntu 18+ with OpenCV. The Open Horizon service yolo utilizes this repository to build Docker containers. Please refer to the Dockerfile for details.

About

OpenYOLO is an open source object detection and classification library written in C with bindings in Python. The library analyzes images and video streams to identify objects and classify them according to a dictionary of eighty (80) entities.

Installation

Outside of use in building the yolo service, run make in the top-level directory to build the darknet/ directory as well as test each version (tiny-v2,tiny-v3,v2,v3); for example (edited for length):

% git clone http://github.com/dcmartin/openyolo
% cd openyolo
% export DARKNET=$(pwd)/darknet
% make
make -C darknet
... lots of lines deleted ...

Usage

Weights for the neural network must be downloaded and made available in the local filesystem; the weights files may be downloaded from the original source:

• tiny, tiny-v2 - http://pjreddie.com/media/files/yolov2-tiny-voc.weights
• tiny-v3 - http://pjreddie.com/media/files/yolov3-tiny.weights
• v2 - https://pjreddie.com/media/files/yolov2.weights
• v3 - https://pjreddie.com/media/files/yolov3.weights

The weights should be downloaded and stored in the top-level directory; for example:

curl -sSL -o ${DARKNET}/yolov2-tiny-voc.weights http://pjreddie.com/media/files/yolov2-tiny-voc.weights

Install CUDA

Install nVidia CUDA drivers and libraries; the information below is for Ubuntu18.04, please refer to the official instructions.

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo add-apt-repository "deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
sudo apt-get update
sudo apt-get -y install cuda

Testing

When the make command completes (successfully), test the build; for example:

% export \
	  DARKNET_WEIGHTS=${OPENYOLO}/yolov2-tiny-voc.weights \
	  DARKNET_CONFIG=${DARKNET}/cfg/yolov2-tiny-voc.cfg \
	  DARKNET_DATA=${DARKNET}/cfg/voc.data \
	  && \
	  ./example/test-yolo.sh
layer     filters    size              input                output
    0 conv     16  3 x 3 / 1   416 x 416 x   3   ->   416 x 416 x  16  0.150 BFLOPs
    1 max          2 x 2 / 2   416 x 416 x  16   ->   208 x 208 x  16
    2 conv     32  3 x 3 / 1   208 x 208 x  16   ->   208 x 208 x  32  0.399 BFLOPs
    3 max          2 x 2 / 2   208 x 208 x  32   ->   104 x 104 x  32
    4 conv     64  3 x 3 / 1   104 x 104 x  32   ->   104 x 104 x  64  0.399 BFLOPs
    5 max          2 x 2 / 2   104 x 104 x  64   ->    52 x  52 x  64
    6 conv    128  3 x 3 / 1    52 x  52 x  64   ->    52 x  52 x 128  0.399 BFLOPs
    7 max          2 x 2 / 2    52 x  52 x 128   ->    26 x  26 x 128
    8 conv    256  3 x 3 / 1    26 x  26 x 128   ->    26 x  26 x 256  0.399 BFLOPs
    9 max          2 x 2 / 2    26 x  26 x 256   ->    13 x  13 x 256
   10 conv    512  3 x 3 / 1    13 x  13 x 256   ->    13 x  13 x 512  0.399 BFLOPs
   11 max          2 x 2 / 1    13 x  13 x 512   ->    13 x  13 x 512
   12 conv   1024  3 x 3 / 1    13 x  13 x 512   ->    13 x  13 x1024  1.595 BFLOPs
   13 conv   1024  3 x 3 / 1    13 x  13 x1024   ->    13 x  13 x1024  3.190 BFLOPs
   14 conv    125  1 x 1 / 1    13 x  13 x1024   ->    13 x  13 x 125  0.043 BFLOPs
   15 detection
mask_scale: Using default '1.000000'
Loading weights from /Volumes/dcmartin/GIT/openyolo/yolov2-tiny-voc.weights...Done!
/Volumes/dcmartin/GIT/openyolo/data/horses.jpg: Predicted in 0.865594 seconds.
cow: 75%

detector.py script

OpenYOLO includes the Python script example/detector.py to run the YOLO algorithm on a specified image; options may be specified for a variety of needs. The Python script takes three arguments (n.b. two and three are optional):

• <JPEG file>- JPEG image file to process
• <config> - may be tiny-v2, tiny-v3, v2, or v3; default: tiny-v2 (optional)
• #.## - minimum value for a positive match; range: [0.0,1.0]; default: 0.25 (optional)

Example

After successfully building the darknet executable use the Python script detector.py to detect entities in the original image; for example:

% ./example/detector.py example/horses.jpg | tee example/horses.json | jq '.'
layer     filters    size              input                output
    0 conv     16  3 x 3 / 1   416 x 416 x   3   ->   416 x 416 x  16  0.150 BFLOPs
    1 max          2 x 2 / 2   416 x 416 x  16   ->   208 x 208 x  16
    2 conv     32  3 x 3 / 1   208 x 208 x  16   ->   208 x 208 x  32  0.399 BFLOPs
    3 max          2 x 2 / 2   208 x 208 x  32   ->   104 x 104 x  32
    4 conv     64  3 x 3 / 1   104 x 104 x  32   ->   104 x 104 x  64  0.399 BFLOPs
    5 max          2 x 2 / 2   104 x 104 x  64   ->    52 x  52 x  64
    6 conv    128  3 x 3 / 1    52 x  52 x  64   ->    52 x  52 x 128  0.399 BFLOPs
    7 max          2 x 2 / 2    52 x  52 x 128   ->    26 x  26 x 128
    8 conv    256  3 x 3 / 1    26 x  26 x 128   ->    26 x  26 x 256  0.399 BFLOPs
    9 max          2 x 2 / 2    26 x  26 x 256   ->    13 x  13 x 256
   10 conv    512  3 x 3 / 1    13 x  13 x 256   ->    13 x  13 x 512  0.399 BFLOPs
   11 max          2 x 2 / 1    13 x  13 x 512   ->    13 x  13 x 512
   12 conv   1024  3 x 3 / 1    13 x  13 x 512   ->    13 x  13 x1024  1.595 BFLOPs
   13 conv    256  1 x 1 / 1    13 x  13 x1024   ->    13 x  13 x 256  0.089 BFLOPs
   14 conv    512  3 x 3 / 1    13 x  13 x 256   ->    13 x  13 x 512  0.399 BFLOPs
   15 conv    255  1 x 1 / 1    13 x  13 x 512   ->    13 x  13 x 255  0.044 BFLOPs
   16 yolo
   17 route  13
   18 conv    128  1 x 1 / 1    13 x  13 x 256   ->    13 x  13 x 128  0.011 BFLOPs
   19 upsample            2x    13 x  13 x 128   ->    26 x  26 x 128
   20 route  19 8
   21 conv    256  3 x 3 / 1    26 x  26 x 384   ->    26 x  26 x 256  1.196 BFLOPs
   22 conv    255  1 x 1 / 1    26 x  26 x 256   ->    26 x  26 x 255  0.088 BFLOPs
   23 yolo
Loading weights from yolov3-tiny.weights...Done!
{
  "file":"data/horses.jpg",
  "meta":"coco.data",
  "net":"cfg/yolov3-tiny.cfg",
  "results":[
    {
      "center":{
        "x":515,
        "y":279
      },
      "confidence":55.9509813785553,
      "entity":"cow",
      "height":143,
      "id":"0",
      "width":180
    },
    {
      "center":{
        "x":85,
        "y":245
      },
      "confidence":55.62857985496521,
      "entity":"horse",
      "height":140,
      "id":"1",
      "width":184
    },
    {
      "center":{
        "x":515,
        "y":282
      },
      "confidence":54.17007803916931,
      "entity":"cow",
      "height":95,
      "id":"2",
      "width":78
    }
  ],
  "threshold":0.5,
  "weights":"yolov3-tiny.weights"
}

yoloanno.sh script

Annotates the original image using the yoloanno.sh script. The script requires ImageMagick and jq software; to install on Debian LINUX:

sudo apt update -qq -y && sudo apt install -qq -y imagemagick jq

Use the the shell script to annotate the image with the expectation of both <example>.jpg and <example>.json exist; for example:

% ./yoloanno.sh example/horses
example/horses-yolo.jpg

The annotated image is stored using the <example>-yolo.json name.

Output

Changelog & Releases

Releases are based on Semantic Versioning, and use the format of MAJOR.MINOR.PATCH. In a nutshell, the version will be incremented based on the following:

• MAJOR: Incompatible or major changes.
• MINOR: Backwards-compatible new features and enhancements.
• PATCH: Backwards-compatible bugfixes and package updates.

Authors & contributors

David C Martin (github@dcmartin.com)