This is an (re-)implementation of DeepLab-ResNet in Pytorch for semantic image segmentation on the PASCAL VOC dataset.
9 July, 2017:
- The training script
train.pyhas been re-written following the original optimisation setup: SGD with momentum, weight decay, learning rate with polynomial decay, different learning rates for different layers, ignoring the 'void' label (255). - The training script with multi-scale inputs
train_msc.pyhas been added: the input is resized to0.5and0.75of the original resolution, and4losses are aggregated: loss on the original resolution, on the0.75resolution, on the0.5resolution, and loss on the all fused outputs. - Evaluation of a single-scale model on the PASCAL VOC validation dataset (using 'SegmentationClassAug') leads to
74.0%mIoU 'VOC12_scenes_20000.pth' without CRF as post-processing step. The evaluation of multi-scale model is in progress.
The DeepLab-ResNet is built on a fully convolutional variant of ResNet-101 with atrous (dilated) convolutions, atrous spatial pyramid pooling, and multi-scale inputs (not implemented here).
The model is trained on a mini-batch of images and corresponding ground truth masks with the softmax classifier at the top. During training, the masks are downsampled to match the size of the output from the network; during inference, to acquire the output of the same size as the input, bilinear upsampling is applied. The final segmentation mask is computed using argmax over the logits.
Optionally, a fully-connected probabilistic graphical model, namely, CRF, can be applied to refine the final predictions.
On the test set of PASCAL VOC, the model achieves 79.7% with CRFs and 76.4% without CRFs of mean intersection-over-union.
For more details on the underlying model please refer to the following paper:
@article{CP2016Deeplab,
title={DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs},
author={Liang-Chieh Chen and George Papandreou and Iasonas Kokkinos and Kevin Murphy and Alan L Yuille},
journal={arXiv:1606.00915},
year={2016}
}
To train the network, one can use the augmented PASCAL VOC 2012 dataset with 10582 images for training and 1449 images for validation.
You can download converted init.caffemodel with extension name .pth here. Besides that, one can also exploit random scaling and mirroring of the inputs during training as a means for data augmentation. For example, to train the model from scratch with random scale and mirroring turned on, simply run:
python train.py --random-mirror --random-scale --gpu 0The single-scale model shows 74.0% mIoU on the Pascal VOC 2012 validation dataset ('SegmentationClassAug'). No post-processing step with CRF is applied.
The following command provides the description of each of the evaluation settings:
python evaluate.py --help