This is my Final Project for the Big Data & Machine Learning Bootcamp I am taking at Core CODE School.
Car view Project: Predicting model for car detection from images and videos.
This is an implementation of Mask R-CNN on Python 3, Keras, and TensorFlow. The model generates bounding boxes for each instance of an object in the image. It's based on Feature Pyramid Network (FPN) and a ResNet101 backbone.
Starting from the dataset "Car Object Detection" from kaggle, we used the Mask R-CNN model to make a predicting model focused on car detection in images and videos.
Then some functions were added in order to make predictions directly with new images and videos.
It's recommended to use a new virtual environment to install this dependencies. This environment must be created with Python 3.6 version.
Use the package manager pip to install the requirements.txt
pip install -r requirements.txtFirst you must consider clone this repository in order to install the correct version of mrcnn==2.1.
In summary, to train the model on your own dataset you'll need to extend two classes:
Config
This class contains the default configuration. Subclass it and modify the attributes you need to change.
Dataset
This class provides a consistent way to work with any dataset.
It allows you to use new datasets for training without having to change
the code of the model. It also supports loading multiple datasets at the
same time, which is useful if the objects you want to detect are not
all available in one dataset.
You can also download pre-trained COCO weights (mask_rcnn_coco.h5) from the Mask_RCNN releases page in order to start with better results in your model training.
With your own model already trained, you can store it in "/model" path and start using the dedicated functions.
This functions have default reading file paths:
videos -> /videos
images -> /images
NOTE: < filename > must be always wrote with extension -> *.jpg, *.mp4 ...
Usage of functions: from terminal and in root path of the repo:
- Play original video
python play_video.py <filename>- Play predicted video
python play_predicted_video.py <filename>- Predict image
python predict_image.py <filename>- Predict video
python predict_video.py <filename>
You can send a pull-request if you want to share some improvements/ideas.
Thanks for this tools and free access info to:
Matterport (Mask R-CNN repository)
Jason Brownlee (very helpful post)