Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/tomfreudenberg/tensorflow-objectdetection
Running example for TensorFlow and object detection
https://github.com/tomfreudenberg/tensorflow-objectdetection
object-detection tensorflow tensorflow-tutorials
Last synced: about 1 month ago
JSON representation
Running example for TensorFlow and object detection
- Host: GitHub
- URL: https://github.com/tomfreudenberg/tensorflow-objectdetection
- Owner: TomFreudenberg
- Created: 2020-05-28T01:02:26.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2020-05-28T01:27:52.000Z (over 4 years ago)
- Last Synced: 2024-10-27T23:44:44.290Z (3 months ago)
- Topics: object-detection, tensorflow, tensorflow-tutorials
- Language: Python
- Homepage: https://www.tensorflow.org/
- Size: 6.84 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# README.md
This is a running walkthru to use TensorFlows [(www.tensorflow.org)](https://www.tensorflow.org/) object detection capabilities.
# Clone tutorial project
```bash
# use your projects folder
cd /home
# clone this repository
git clone https://github.com/TomFreudenberg/TensorFlow-ObjectDetection.git tf-tutorial
cd tf-tutorial
```
# Run tutorial
If all tools and packages installed, you may run the object_detection example:
```bash
cd src/tutorial
python image_object_detection.py
```
# Setup environment
```bash
# prepare a virtual environment path
mkdir -p venv/bin
# install packages
pip install tensorflow
pip install Cython
pip install contextlib2
pip install pillow
pip install lxml
pip install jupyter
pip install matplotlib
```
### Download latest Protobuf
Download the latest Protobuf version for your OS and extract it.
* https://github.com/google/protobuf/releases
##### OSX binary (e.g.)
```bash
cd venv
curl -L -O https://github.com/protocolbuffers/protobuf/releases/download/v3.12.2/protoc-3.12.2-osx-x86_64.zip
unzip protoc-3*.zip
cd ..
```
### Download TensorFlow models from github
Clone or download TensorFlow’s Model from Github. The downloaded archive must be extracted as "models" folder.
* https://github.com/tensorflow/models
```bash
mkdir -p src/tensorflow
cd src/tensorflow
curl -L -O https://github.com/tensorflow/models/archive/master.zip
unzip master.zip
mv models-master models
cd ../..
```
Patch models to be compatible with TensorFlow 2 and newer tools
```bash
(
echo '--- input_reader.proto'
echo '+++ input_reader.proto'
echo '@@ -4 +4 @@'
echo '-import "object_detection/protos/image_resizer.proto";'
echo '+// import "object_detection/protos/image_resizer.proto";'
) | patch -s src/tensorflow/models/research/object_detection/protos/input_reader.proto
(
echo '--- ops.py'
echo '+++ ops.py'
echo '@@ -27,2 +27,2 @@'
echo '-import tensorflow.compat.v1 as tf'
echo '-import tf_slim as slim'
echo '+import tensorflow as tf'
echo '+'
echo '@@ -846 +846 @@'
echo '- box_ind=tf.range(num_boxes),'
echo '+ box_indices=tf.range(num_boxes),'
) | patch -s src/tensorflow/models/research/object_detection/utils/ops.py
```
### Run protoc to create .py from proto files
Open the research folder and run protobuf:
```bash
cd src/tensorflow/models/research/
../../../../venv/bin/protoc object_detection/protos/*.proto --python_out=.
cd ../../../..
```
Check if this had worked, while looking at the protos folder. Every proto file should have a python file created.
# Credits
* Tutorial inspired from: https://www.edureka.co/blog/tensorflow-object-detection-tutorial/