This repository is a tutorial for how to use TensorFlow Object Detection
https://developer.nvidia.com/cuda-90-download-archive (CUDA Toolkit 9.0)
https://developer.nvidia.com/rdp/cudnn-download (cuDNN v7.0.5 (Dec 5, 2017), for CUDA 9.0)
https://www.anaconda.com/download/
In the command terminal that pops up, create a new virtual environment called “tensorflow1” by issuing the following command:
C:\> conda create -n tensorflow1 pip python=3.5
Then, activate the environment by issuing:
C:\> activate tensorflow1
Install tensorflow-gpu in this environment by issuing:
(tensorflow1) C:\> pip install --ignore-installed --upgrade tensorflow-gpu
Install the other necessary packages by issuing the following commands:
(tensorflow1) C:\> pip install pillow
(tensorflow1) C:\> pip install Cython
(tensorflow1) C:\> pip install git+https://github.com/philferriere/cocoapi.git#egg=pycocotools^&subdirectory=PythonAPI
(tensorflow1) C:\> pip install pandas
(tensorflow1) C:\> pip install lxml
(tensorflow1) C:\> pip install contextlib2
(tensorflow1) C:\> pip install opencv-python
https://github.com/philferriere/cocoapi to install pycocotools
You can choose one method to create PYTHONPATH
(tensorflow1) C:\> set PYTHONPATH=C:\tensorflow1\models;C:\tensorflow1\models\research\slim
Or in Search, search for and then select: System (Control Panel)
Click the System and then Advanced system settings link.
Click Environment Variables. In the section System Variables, click new.
Name: PYTHONPATH
Value: C:\tensorflow1\models;C:\tensorflow1\models\research\slim
Click OK. Close all remaining windows by clicking OK.
You can download it here Protocol Buffers v3.0.2
(tensorflow1) C:\> cd C:\tensorflow1\models\research
(tensorflow1) C:\tensorflow1\models\research> C:/Users/name/Desktop/bin/protoc object_detection/protos/*.proto --python_out=.
Finally, run the following commands from the C:\tensorflow1\models\research directory:
(tensorflow1) C:\tensorflow1\models\research> python setup.py build
(tensorflow1) C:\tensorflow1\models\research> python setup.py install
We'll use faster_rcnn_inception_v2_coco models.
We can download models here faster_rcnn_inception_v2_coco
Once downloaded, unzip the folder in /object_detection. The folder must have file frozen_inference_graph.pb (models)
We use for example faster_rcnn_inception_v2_coco models
Save folder faster_rcnn_inception_v2_coco_2018_01_28 in C:\tensorflow1\models\research\object_detection
The label map tells the trainer what each object is by defining a mapping of class names to class ID numbers. Use a text editor to create a new file and save it as labelmap.pbtxt in the C:\tensorflow1\models\research\object_detection\output folder, will look like.
item {
id: 1
name: 'raccoon'
}
Finally, the object detection training pipeline must be configured. It defines which model and what parameters will be used for training.
Navigate to C:\tensorflow1\models\research\object_detection\samples\configs and copy the faster_rcnn_inception_v2_pets.config file into the \object_detection\output directory and change "PATH_TO_BE_CONFIGURED",num_classe,num_examples for correct data, like here
We have to save our image in imagenes/
First, the image .xml data will be used to create TFRecord files containing all the data for the images. From the \object_detection folder, issue the following command in the Anaconda command prompt:
(tensorflow1) C:\tensorflow1\models\research\object_detection> python create_pet_tf_record.py --data_dir=imagenes --output_dir=output --label_map_path=output/labelmap.pbtxt
(tensorflow1) C:\tensorflow1\models\research\object_detection> python model_main.py --pipeline_config_path=output/faster_rcnn_inception_v2_pets.config --model_dir=trained_model --alsologtostderr
If you want to see all the graphs, run this command
(tensorflow1) C:\tensorflow1\models\research\object_detection> tensorboard --logdir=trained_model
Now that training is complete, the last step is to generate the frozen inference graph (.pb file).
“XXX” in “model.ckpt-XXX” should be replaced with the highest-numbered .ckpt file in the training folder:
(tensorflow1) C:\tensorflow1\models\research\object_detection> python export_inference_graph.py --pipeline_config_path=output/faster_rcnn_inception_v2_pets.config --trained_checkpoint_prefix=trained_model/model.ckpt-XXX --output_directory=inference_graph
the file frozen_inference_graph.pb has been generated, this is your models
Now you have your models (frozen_inference_graph.pb).
To run it
(tensorflow1) C:\tensorflow1\models\research\object_detection> python Object_detection_webcam.py
Output
