Part-level Dataset Available for Augmentation

Sources

Single Instance

• Image Classification - 32 Classes - Fourniture: About 10K

Complicated Scene

• Real
  • Indoor Training Set (ITS) [RESIDE-Standard]: 1.4K
  • MIT Indoor Scenes: 15K
• Synthetic
  • InteriorVerse

chair_dataset

• image_1~image_3000: kaggle furniture image dataset
• image_3001~image_4887: DeepFurniture

desk_dataset

• image_1~image_700: pix3d

home_appliance_dataset

• image_1~image_3000: kaggle furniture image dataset fridge only
• image_3001~image_3429: DeepFurniture home-appliance category

shelves_dataset

• image_1~image_3000: kaggle furniture image dataset
• image_3001~image_3243: pix3d wardrobe category
• image_3244~image_3604: pix3d bookcase category

sofa_dataset

• image_1~image_1947: pix3d
• image_1498~image_3888: DeepFurniture

table_dataset

• image_1~image_3000: kaggle furniture image dataset
• image_3001~image_4870: pix3d
• image_4871~image_7293: DeepFurniture

tool_dataset

• image_1~image_115: pix3d
• image_116~image_1441:kaggle mechanical tool dataset hammer
• image_1442~image_1812:kaggle mechanical tool dataset plier
• image_1813~image_3138:kaggle mechanical tool dataset screw driver
• image_3139~image_4469:kaggle mechanical tool dataset wrench

tv_dataset

• image_1~image_3000: kaggle furniture image dataset

PSR Dataset

Cabinet (Processed)

from shelves_dataset # 1 3000 images -> 1081 train & 163 val samples

Desk (Processed)

from desk_dataset 700 images -> 2599 train & 665 val samples

Tool (Processed)

from tool_dataset (simplified) 1335 images -> 2999 train & 766 val samples

Furniture (Processed)

from 130k Images/furniture 1983 images -> 5736 train & 1434 val samples

current total:

• train: 12,415-686(bg)=11729
• val: 3,028-182(bg)=2846