https://github.com/zfturbo/keras-mask-rcnn-for-open-images-2019-instance-segmentation

Code and pre-trained models for Instance Segmentation track in Open Images Dataset
https://github.com/zfturbo/keras-mask-rcnn-for-open-images-2019-instance-segmentation

keras mask-rcnn open-images-dataset pretrained-models

Last synced: 2 months ago
JSON representation

Code and pre-trained models for Instance Segmentation track in Open Images Dataset

• Host: GitHub
• URL: https://github.com/zfturbo/keras-mask-rcnn-for-open-images-2019-instance-segmentation
• Owner: ZFTurbo
• Created: 2019-10-25T08:05:11.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2019-10-25T16:33:05.000Z (about 5 years ago)
• Last Synced: 2024-03-20T23:10:33.453Z (10 months ago)
• Topics: keras, mask-rcnn, open-images-dataset, pretrained-models
• Language: Python
• Homepage:
• Size: 182 KB
• Stars: 57
• Watchers: 4
• Forks: 16
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## Keras Mask R-CNN for Open Images Challenge 2019: Instance Segmentation

Repository contains Mask R-CNN models which were trained on Open Images Dataset during Kaggle competition: 

https://www.kaggle.com/c/open-images-2019-instance-segmentation/leaderboard 

Repository contains the following:

* Pre-trained Mask R-CNN models (ResNet50, ResNet101 and ResNet152 backbones)

* Example code to get predictions with these models for any set of images

* Code to train model based on Keras Mask R-CNN and OID dataset 

## Requirements

Python 3.\*, Keras 2.\*, [keras-maskrcnn 0.2.2](https://github.com/fizyr/keras-maskrcnn), cv2, numpy, pandas

## Pretrained models

There are 3 Mask R-CNN models based on ResNet50, ResNet101 and ResNet152 for [300 classes](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/blob/master/data_segmentation/challenge-2019-classes-description-segmentable.csv). 

| Backbone | Image Size (px) | Model | Small validation mAP | LB (Public) |

| --- | --- | --- | --- | --- | 

| ResNet50 | 800 - 1024 | [521 MB](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/releases/download/v1.0/mask_rcnn_resnet50_oid_v1.0.h5) | 0.5745 | 0.4259 |

| ResNet101 | 800 - 1024 | [739 MB](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/releases/download/v1.0/mask_rcnn_resnet101_oid_v1.0.h5) | 0.5917 | 0.4345 |

| ResNet152 |  800 - 1024 | [918 MB](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/releases/download/v1.0/mask_rcnn_resnet152_oid_v1.0.h5) | 0.5899 | 0.4404 |

* Model - can be used to resume training or can be used as pretrain for your own instance segmentation model

## Inference 

Simple example can be found here: [inference_example.py](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/blob/master/inference_example.py)

![Example of predictions](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/blob/master/img/mask_rcnn_prediction_example.jpg)

## Training

For training you need to download OID dataset (~500 GB images): https://storage.googleapis.com/openimages/web/download.html

You need all images, all masks and all CSV-files related to Instance Segmentation track. 

Then run script (change parameters and file locations at the bottom of script):

* [training/train_maskrcnn.py](https://github.com/ZFTurbo/Keras-Mask-RCNN-for-Open-Images-2019-Instance-Segmentation/blob/master/training/train_maskrcnn.py)