Home

Awesome

COCO_2018-Stuff-Segmentation-Challenge

COCO_2018-Stuff-Segmentation-Challenge What is COCO? COCO is large-scale object detection, segmentation, and captioning dataset. The COCO Stuff Segmentation Task is designed to push the state of the art in semantic segmentation of stuff classes. Whereas the object detection task addresses thing classes (person, car, elephant), this task focuses on stuff classes (grass, wall, sky). For full details of the stuff segmentation task please see the stuff evaluation page. Note: the newly introduced panoptic segmentation task addresses recognition of both things and stuff classes simultaneously. This project implicates schematic segmentation, where raw RGB images are processed into the model and pixel masked images are given out as an output.

Internship project in Bennett University under Leadinginadi.ai

Getting Started

Problem Statement :-<br/> To perform Semantic Segmentation of Stuff classes.The COCO Stuff Segmentation Task is designed to push the state of the art in semantic segmentation of stuff classes.

Prerequisites and installing

you can run script.sh in the scripts folder to install all prequisites. selective prequisites are given below. pip install tensorflow-gpu<br/> pip install tqdm<br/> pip install keras<br/> pip install keras-segmentation<br/>

Dataset format

You need to make two folders<br/>

Images Folder - For all the training images<br/> Annotations Folder - For the corresponding ground truth segmentation images<br/> The filenames of the annotation images should be same as the filenames of the RGB images.<br/>

Usage via command line

Visualizing the prepared data

python -m keras_segmentation verify_dataset \
 --images_path="dataset1/images_prepped_train/" \
 --segs_path="dataset1/annotations_prepped_train/"  \
 --n_classes=50
python -m keras_segmentation visualize_dataset \
 --images_path="dataset1/images_prepped_train/" \
 --segs_path="dataset1/annotations_prepped_train/"  \
 --n_classes=50

Training the Model

python -m keras_segmentation train \
--checkpoints_path="path_to_checkpoints" \
--train_images="dataset1/images_prepped_train/" \
--train_annotations="dataset1/annotations_prepped_train/" \
--val_images="dataset1/images_prepped_test/" \
--val_annotations="dataset1/annotations_prepped_test/" \
--n_classes=300 \
--input_height=320 \
--input_width=640 \
--model_name="pspnet"

Getting the predictions

python -m keras_segmentation predict \
 --checkpoints_path="path_to_checkpoints" \
 --input_path="dataset1/images_prepped_test/" \
 --output_path="path_to_predictions"

References

  1. https://github.com/divamgupta/image-segmentation-keras<br/>
  2. https://github.com/GeorgeSeif/Semantic-Segmentation-Suite<br/>
  3. https://github.com/aurora95/Keras-FCN<br/>
  4. http://cocodataset.org/#stuff-2018<br/>
  5. https://arxiv.org/abs/1612.03716<br/>