Skip to content

Latest commit

 

History

History
32 lines (18 loc) · 1.79 KB

DATA.md

File metadata and controls

32 lines (18 loc) · 1.79 KB

Data Preparation

We use mmsegmentation for semantic segmentation testing, evaluation and visualization on Pascal VOC, Pascal Context and COCO datasets. You do not need to construct a tree-like data structure within the codebase. To specify the root of each folder, you need to modify the data_root attribute in each config file. All datasets include the "background" class for mIoU evaluation.

Semantic Segmentation

Pascal VOC

Please follow the MMSegmentation Pascal VOC Preparation instructions to download and setup the Pascal VOC dataset.

Pascal Context

Please refer to the MMSegmentation Pascal Context Preparation instructions to download and setup the Pascal Context dataset.

COCO Object

We follow GroupViT to prepare the data of COCO Object.

ADE-150 and ADE-847

We follow FC-CLIP to prepare the data of ADE-150 and ADE-847.

Referring Segmentation

Ref-COCO, RefCOCO+ and RefCOCOg

  1. Follow instructions in the refer directory to set up subdirectories and download annotations.

  2. Download images from COCO. Please use the first downloading link 2014 Train images [83K/13GB], and extract the downloaded train_2014.zip file to ./refer/data/images/mscoco/images.