Skip to content

a tensorflow implement of refinenet ,RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation

Notifications You must be signed in to change notification settings

eragonruan/refinenet-image-segmentation

Repository files navigation

refinenet

a tensorflow implement of refinenet. RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation

Introduction

this is a tensorflow implement of refinenet discribed in arxiv:1611.06612.I have not finished it yet, this is just a demo, but the model is already able to work.

prepare

  • download the pretrain model of resnet_v1_101.ckpt, you can download it from here
  • download the pascal voc dataset
  • some dependence like cv2, numpy and etc. recommend to install Anaconda

training

  • first, run convert_pascal_voc_to_tfrecords.py to convert training data into .tfrecords, Or you can use the tfrecord I converted In BaiduYun.Currently, I only use the pascal voc 2012 for training.
  • second, run python RefineNet/multi_gpu_train.py, also, you can change some hyper parameters in this file, like the batch size.

eval

  • if you have already got a model, or just download the model I trained on pascal voc.model.
  • put images in demo/ and run python RefineNet/demo.py

roadmap

  • python2/3 compatibility
  • Complete realization of refinenet model
  • test on pascal voc, give the IoU result
  • training on other datasets

some result

About

a tensorflow implement of refinenet ,RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages