Tutorial on how to create your own object detection dataset and train using TensorFlow's API
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 5 commits ahead of wagonhelm:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
ApproachingCars.ipynb
README.md
generate_tfrecord.py
json_to_csv.py
requirements.txt
split_labels.ipynb
xml_to_csv.py

README.md

This repo was forked from TensorFlow Object Detection API Tutorial found here

https://github.com/wagonhelm/TF_ObjectDetection_API:

Required Packages

See my blog post to learn about what I did: http://www.bekcunning.com/blog/car-back

FILES

  1. ApproachingCars.ipynb Contains the steps required to run the car detector and construct a video from the results.

  2. json_to_csv.py and xml_to_csv.py Contain code to convert the image attribute files to csv which will be used to generate tfrecords

  3. split_labels.ipynb Creates test and training data from our complete set of images

  4. generate_tfrecord.py python code required to generate tfrecords used during training of the model

  5. data/ this folder contains the moodel config and file for labelling the desired classes.