{"id":1605,"date":"2023-11-05T21:53:19","date_gmt":"2023-11-05T12:53:19","guid":{"rendered":"http:\/\/www.ecobioinfo.com\/?p=1605"},"modified":"2023-11-11T15:20:40","modified_gmt":"2023-11-11T06:20:40","slug":"%e6%b7%b1%e5%b1%a4%e5%ad%a6%e7%bf%92%e3%80%80windwos10-%e3%81%ab-tensorflow-object-detection-api-%e3%82%92%e3%82%a4%e3%83%b3%e3%82%b9%e3%83%88%e3%83%bc%e3%83%ab","status":"publish","type":"post","link":"https:\/\/www.ecobioinfo.com\/?p=1605","title":{"rendered":"\u6df1\u5c64\u5b66\u7fd2\u3000Windwos10 \u306b TensorFlow Object Detection API \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb"},"content":{"rendered":"

\u524d\u56de Ubuntu \u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/a>\u3057\u305f\u306e\u3067\u3001
\n\u4eca\u5ea6\u306f Tensorflow\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u691c\u51faAPI\u3092\u958b\u767a\u74b0\u5883\u306e\u7121\u3044 Windows10 \u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u307f\u307e\u3057\u305f\u3002
\n\u958b\u767a\u74b0\u5883\u306e\u7121\u3044\u4e00\u822c\u5411\u3051\u306eWindows\u306a\u306e\u3067\u3001\u5f8c\u8ff0\u306e\u901a\u308a Git \u3084 VisualStudio \u306a\u3069\u306e\u958b\u767a\u30c4\u30fc\u30eb\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002
\n\uff08GPU\u8a2d\u5b9a\u306b\u3064\u3044\u3066\u306f\u4eca\u56de\u306f\u3068\u308a\u3042\u3048\u305a\u7701\u7565\u3057\u307e\u3059\u3002\uff09<\/p>\n

 
\n\uff1c\uff1c\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\uff1e\uff1e<\/p>\n

\u57fa\u672c\u7684\u306b\u306f\u30ea\u30f3\u30af\u5148\u306e\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306b\u6cbf\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n

TensorFlow 2 Object Detection API tutorial
\nhttps:\/\/tensorflow-object-detection-api-tutorial.readthedocs.io\/en\/latest\/install.html<\/a><\/p>\n

 
\n1. Anaconda Python 3.8 (Optional)
\n1.1 Anaconda Python 3.8 \u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/p>\n

Anaconda \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u307e\u3059\u3002<\/p>\n

https:\/\/www.anaconda.com\/download<\/a><\/p>\n

Anaconda3-2023.09-0-Windows-x86_64.exe \u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3001\u5f8c\u306f\u8d77\u52d5\u3057\u3066GUI\u30a4\u30f3\u30b9\u30c8\u30fc\u30e9\u306e\u901a\u308a\u3002<\/p>\n

conda\u30d0\u30fc\u30b8\u30e7\u30f3\u78ba\u8a8d
\n(base) C:\\(userdir)>conda -V
\nconda 23.7.4<\/p>\n

1.2 Create a new Anaconda virtual environment<\/p>\n

Anaconda prompt\u3088\u308a\u4eee\u60f3\u74b0\u5883 tfod \u3092\u4f5c\u6210\u3002python \u306f3.9 \u3068\u3057\u307e\u3057\u305f\u3002<\/p>\n

(base) C:\\(userdir)>conda create -n tfod pip python=3.9<\/p>\n

\u4eee\u60f3\u74b0\u5883\u30a2\u30af\u30c6\u30a3\u30d6
\n(base) C:\\(userdir)\\anaconda3\\envs\\tfod>conda activate tfod
\n(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod><\/p>\n

\u30d0\u30fc\u30b8\u30e7\u30f3\u78ba\u8a8d
\n(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod>python -V
\nPython 3.9.18<\/p>\n

\u4eee\u60f3\u74b0\u5883\u306b Python 3.9.18 \u304c\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n

2. TensorFlow Installation
\n2.1 Install the TensorFlow PIP package<\/p>\n

PIP\u3067TensorFlow\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/p>\n

(tfod) C:\\(usedir)\\anaconda3\\envs\\tfod>pip install –ignore-installed –upgrade tensorflow==2.5.0
\n\u30fb\u30fb\u30fb
\nCollecting tensorflow==2.5.0
\n Downloading tensorflow-2.5.0-cp39-cp39-win_amd64.whl (422.6 MB)
\n\u30fb\u30fb\u30fb<\/p>\n

3. GPU Support (Optional)<\/p>\n

GPU\u8a2d\u5b9a\u306b\u3064\u3044\u3066\u306f\u4eca\u56de\u306f\u3068\u308a\u3042\u3048\u305a\u7701\u7565\u3057\u307e\u3059\u3002<\/p>\n

4. TensorFlow Object Detection API Installation
\nTensorFlow Object Detection API\u3000\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3002<\/p>\n

4.1 Downloading the TensorFlow Model Garden<\/p>\n

Zip\u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9
\nhttps:\/\/github.com\/tensorflow\/models\/archive\/master.zip<\/a><\/p>\n

\u5c55\u958b\u3057\u3066\u3001Tensorflow\\models\u30d5\u30a9\u30eb\u30c0\u3092\u4f5c\u6210\u3057\u3066models-master\u306e\u5185\u5bb9\u3092\u30b3\u30d4\u30fc\u3057\u307e\u3057\u305f\u3002<\/p>\n

4.2 Protobuf Installation\/Compilation<\/p>\n

Protobuf\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3002<\/p>\n

\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u30da\u30fc\u30b8
\nhttps:\/\/github.com\/protocolbuffers\/protobuf\/releases<\/a><\/p>\n

Windwos\u7528\u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9
\nprotoc-24.3-win64.zip<\/p>\n

C:\\(userdir)\\protobuf\u306b\u5c55\u958b\u3001\u74b0\u5883\u5909\u6570\u306b\u30d1\u30b9\u8a2d\u5b9a\u3001\u4eee\u60f3\u74b0\u5883\u518d\u8d77\u52d5<\/p>\n

\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u691c\u51faAPI\u3092\u30b3\u30d4\u30fc\u3057\u305f\u30c7\u30a3\u30ec\u30af\u30c8\u30ea\u306e Tensorflow\/models\/research\/\u306b\u79fb\u52d5\u3057\u3066\u3001
\n\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306b\u66f8\u3044\u3066\u3042\u308b\u30b3\u30de\u30f3\u30c9\u5b9f\u884c\u3002<\/p>\n

(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod\\models\\research>protoc object_detection\/protos\/*.proto –python_out=.<\/p>\n

4.3 COCO API installation
\nCOCO API \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3002<\/p>\n

\u307e\u305a cython \u304b\u3089\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/p>\n

(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod\\pip install cython
\nInstalling collected packages: cython
\nSuccessfully installed cython-3.0.2<\/p>\n

\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306b\u306f\u3001VC++2015 Build tools \u304c\u5fc5\u8981\u3068\u66f8\u3044\u3066\u3042\u308a\u307e\u3059\u3002
\n“Note that, according to the package\u2019s instructions, Visual C++ 2015 build tools must be installed and on your path. If they are not, make sure to install them from here.”<\/p>\n

\u4e00\u822c\u5411\u3051 Winodows PC \u306a\u306e\u3067\u5fc5\u8981\u306a\u958b\u767a\u74b0\u5883\u304c\u3042\u308a\u307e\u305b\u3093\u3002
\n\u307e\u305a\u3001GIT\u304b\u3089\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u307e\u3057\u305f\u3002<\/p>\n

\u53c2\u8003
\nhttps:\/\/www.curict.com\/item\/60\/60bfe0e.html<\/a>
\n\u516c\u5f0f
\nhttps:\/\/gitforwindows.org\/<\/a><\/p>\n

Windows64bit\u7248 Git-2.42.0.2-64-bit.exe \u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u3066\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3002<\/p>\n

VC++2015\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/p>\n

\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306e”from here”\u306e\u30ea\u30f3\u30af\u5148\u3067\u306f\u6b63\u5e38\u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3067\u304d\u306a\u304b\u3063\u305f\u306e\u3067\u3001
\nWeb\u30da\u30fc\u30b8\u3092\u53c2\u8003\u306b\u3057\u3066\u3001vs_BuildTools.exe, visualcppbuildtools_full.exe
\n\u306a\u3069\u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u3066\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u307f\u307e\u3057\u305f\u3002<\/p>\n

https:\/\/visualstudio.microsoft.com\/ja\/visual-cpp-build-tools\/<\/a><\/p>\n

https:\/\/go.microsoft.com\/fwlink\/?LinkId=691126<\/a><\/p>\n

\u6e96\u5099\u304c\u3067\u304d\u307e\u3057\u305f\u306e COCOAPI \u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3092\u3057\u307e\u3059\u3002<\/p>\n

(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod>pip install git+https:\/\/github.com\/philferriere\/cocoapi.git#subdirectory=PythonAPI
\n\u30fb\u30fb\u30fb
\nCollecting git+https:\/\/github.com\/philferriere\/cocoapi.git#subdirectory=PythonAPI
\n\u30fb\u30fb\u30fb\uff08\u4e2d\u7565\uff09
\nSuccessfully built pycocotools
\nInstalling collected packages: pycocotools
\nSuccessfully installed pycocotools-2.0<\/p>\n

\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3055\u308c\u305f\u3088\u3046\u3067\u3059\u3002<\/p>\n

4.4 Install the Object Detection API<\/p>\n

\u3088\u3046\u3084\u304f\u6e96\u5099\u304c\u3067\u304d\u307e\u3057\u305f\u306e\u3067\u3001\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u691c\u51faAPI\u306e\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3067\u3059\u3002<\/p>\n

(tfod) C:\\(userdir)\\anaconda3\\envs\\tfod\\Tensorflow\\models\\research>copy object_detection\\packages\\tf2\\setup.py .<\/p>\n

(tfod) C:\\(userdir)\\\\anaconda3\\envs\\tfod\\Tensorflow\\models\\research>python -m pip install .<\/p>\n

(tfod) C:\\(userdir)\\\\anaconda3\\envs\\tfod\\Tensorflow\\models\\research>python object_detection\/builders\/model_builder_tf2_test.py<\/p>\n

\u30fb\u30fb\u30fb\uff08\u30e1\u30c3\u30bb\u30fc\u30b8\u4e00\u90e8\u629c\u7c8b\uff09
\n2023-10-07 07:29:00.747795: W tensorflow\/stream_executor\/platform\/default\/dso_loader.cc:64] Could not load dynamic library ‘cudart64_110.dll’; dlerror: cudart64_110.dll not found
\n2023-10-07 07:29:00.747947: I tensorflow\/stream_executor\/cuda\/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
\nTraceback (most recent call last):
\n File “C:\\(userdir)\\anaconda3\\envs\\tfod\\Tensorflow\\models\\research\\object_detection\\builders\\model_builder_tf2_test.py”, line 24, in
\n from object_detection.builders import model_builder
\n File “C:\\(userdir)\\anaconda3\\envs\\tfod\\lib\\site-packages\\object_detection\\builders\\model_builder.py”, line 23, in
\n from object_detection.builders import anchor_generator_builder
\n File “C:\\(userdir)\\anaconda3\\envs\\tfod\\lib\\site-packages\\object_detection\\builders\\anchor_generator_builder.py”, line 26, in
\n from object_detection.protos import anchor_generator_pb2
\nImportError: cannot import name ‘anchor_generator_pb2’ from ‘object_detection.protos’ (C:\\(userdir)\\envs\\tfod\\lib\\site-packages\\object_detection\\protos\\__init__.py)<\/p>\n

GPU\u306e\u30a6\u30a9\u30fc\u30cb\u30f3\u30b0\u306fGPU\u74b0\u5883\u3092\u8a2d\u5b9a\u3057\u3066\u306a\u3044\u306e\u3067\u826f\u3044\u306e\u3067\u3059\u304c\u3001
\n\u6700\u5f8c\u306b\u3001\u30a8\u30e9\u30fc\u304c\u51fa\u307e\u3057\u305f\u3002<\/p>\n

\u4e0b\u8a18\u3092\u53c2\u8003\u306b\u3057\u3066\u5bfe\u51e6\u3057\u3066\u307f\u307e\u3057\u305f\u3002
\nhttps:\/\/qiita.com\/kappanda\/items\/697ba3c6625774d3414c<\/a><\/p>\n

https:\/\/github.com\/tensorflow\/models\/issues\/1595<\/a><\/p>\n

https:\/\/qiita.com\/johnpapapa\/items\/e7e06f2480251102d16d<\/a><\/p>\n

\u30c1\u30e5\u30fc\u30c8\u30ea\u30a2\u30eb\u306e\u30ea\u30f3\u30af\u5148\u304b\u3089
\nhttps:\/\/github.com\/protocolbuffers\/protobuf\/releases<\/a>
\nprotobuf-24.3 \u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u3066\u5c55\u958b\u3002<\/p>\n

protobuf-24.3\\python\\google\\protobuf\\internal
\n\u304b\u3089
\nC:\\(userdir)\\anaconda3\\envs\\tfod\\Lib\\site-packages\\google\\protobuf\\internal<\/p>\n

\u306b\u8db3\u308a\u306a\u3044\u30bd\u30fc\u30b9 builder.py \u3092\u30b3\u30d4\u30fc\u3057\u3066\u89e3\u6d88\u3057\u307e\u3057\u305f\u3002
\nUbuntu\u306b\u304f\u3089\u3079\u3066\u554f\u984c\u304c\u591a\u304f\u3066\u82e6\u52b4\u3057\u307e\u3057\u305f\u304c\u3001\u8a66\u884c\u932f\u8aa4\u3057\u3066\u3001\u306a\u3093\u3068\u304b\u30a8\u30e9\u30fc\u306f\u89e3\u6d88\u3055\u308c\u307e\u3057\u305f\u3002<\/p>\n

 
\n 
\n\uff1c\uff1c\u30b5\u30f3\u30d7\u30eb\u5b9f\u884c\uff1e\uff1e<\/p>\n

\u30b5\u30f3\u30d7\u30eb\u30bd\u30fc\u30b9\u3092\u30c0\u30a6\u30f3\u30ed\u30fc\u30c9\u3057\u3066\u30b3\u30de\u30f3\u30c9\u30d7\u30ed\u30f3\u30d7\u30c8\u304b\u3089\u5b9f\u884c\u3057\u3066\u307f\u307e\u3057\u305f\u3002<\/p>\n

https:\/\/tensorflow-object-detection-api-tutorial.readthedocs.io\/en\/latest\/auto_examples\/index.html#examplesplot_object_detection_saved_model.py<\/a>
\n\u30b5\u30f3\u30d7\u30eb\u30d7\u30ed\u30b0\u30e9\u30e0\u306f\u3053\u306e\u307e\u307e\u3060\u3068 Anaconda \u30b3\u30de\u30f3\u30c9\u30d7\u30ed\u30f3\u30d7\u30c8\u306e\u5b9f\u884c\u3067\u306f\u753b\u50cf\u30a6\u30a3\u30f3\u30c9\u30a6\u304c\u8868\u793a\u3055\u308c\u306a\u3044\u306e\u3067\u3001
\nplt.imshow \u306e\u4ee3\u308f\u308a\u306b OpenCV \u3067\u8868\u793a\u3059\u308b\u3088\u3046\u306b\u4fee\u6b63\u3057\u307e\u3057\u305f\u3002<\/p>\n

plot_object_detection_saved_model.py \u306e\u5909\u66f4\u7b87\u6240
\n \u4e00\u90e8\u629c\u7c8b ———————–
\nimport cv2
\n\uff08\u4e2d\u7565\uff09
\n plt.figure()
\n plt.imshow(image_np_with_detections)<\/p>\n

#\u2193\u3053\u308c\u3092\u8ffd\u52a0
\n orgname = os.path.splitext(os.path.basename(image_path))[0]
\n cv2.namedWindow(orgname)
\n cv2.imshow(orgname, cv2.cvtColor(image_np_with_detections, cv2.COLOR_RGB2BGR))
\n #\u2191\u3053\u308c\u3092\u8ffd\u52a0
\nplt.show()<\/p>\n

#\u2193\u3053\u308c\u3092\u8ffd\u52a0
\ncv2.waitKey(0)
\n#\u2191\u3053\u308c\u3092\u8ffd\u52a0
\n————————————<\/p>\n

\u5b9f\u884c\u7d50\u679c<\/p>\n

\"\u30ad\u30e3\u30d7\u30c1\u30e3\"<\/p>\n

\u3044\u308d\u3044\u308d\u3068\u554f\u984c\u3042\u308a\u307e\u3057\u305f\u304c\u3001\u306a\u3093\u3068\u304b\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308a\u307e\u3057\u305f\u3002<\/p>\n

\uff082023\/11\/04 Shin Onda\uff09<\/p>\n","protected":false},"excerpt":{"rendered":"\u524d\u56de Ubuntu \u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u305f\u306e\u3067\u3001 \u4eca\u5ea6\u306f Tensorflow\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u691c\u51faAPI\u3092\u958b\u767a\u74b0\u5883\u306e\u7121\u3044 Windows10 \u306b\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u3066\u307f\u307e\u3057\u305f\u3002 \u958b\u767a\u74b0\u5883\u306e\u7121\u3044\u4e00\u822c\u5411\u3051\u306eWindows\u306a\u306e\u3067\u3001\u5f8c\u8ff0\u306e\u901a\u308a …","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[],"class_list":["post-1605","post","type-post","status-publish","format-standard","hentry","category-ai_ecobio"],"_links":{"self":[{"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/posts\/1605","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1605"}],"version-history":[{"count":18,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/posts\/1605\/revisions"}],"predecessor-version":[{"id":1625,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=\/wp\/v2\/posts\/1605\/revisions\/1625"}],"wp:attachment":[{"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1605"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1605"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.ecobioinfo.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1605"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}