Image Dataset for Resale Item Recognition & Price Estimation

A second hand item image dataset of furniture, appliances, and household items for training resale price estimation models, depth camera calibration, and household item recognition systems.

Browse catalog Request custom dataset
Home / Use Cases / Resale Item Recognition

Built for resale platforms (Gumtree, Facebook Marketplace, eBay) building automated price suggestion, waste collection and logistics companies using depth cameras for item recognition and dimension estimation, insurance and valuation companies estimating household contents from photos, and recycling and reverse logistics companies classifying bulky waste items.

Problem statement

Training a model to recognise and value second-hand items requires images of those items in real-world conditions — not just clean studio product shots. A sofa in a living room, a bed frame in a bedroom, a wardrobe against a wall. These items need to be recognisable from the angle a seller or waste collector would photograph them: often at an angle, in variable lighting, partially obscured. This is what makes ImageHub a practical furniture recognition dataset rather than a purely studio-clean one.

Additionally, to validate size and weight estimates from depth cameras, reference measurements for standard item types (beds 135cm × 190cm, standard sofa 220cm × 90cm × 85cm) are needed as ground truth.

How ImageHub solves it

As used furniture image training data goes, this is closer to real-world conditions than pure studio datasets:

  • Furniture, home furnishings, and appliance images from lifestyle photography settings (not just white-background studio shots) — closer to real-world conditions than pure studio datasets.
  • Multiple angles per product — front, side, top-down — supports multi-angle recognition models.
  • Category labels (sofas, beds, chairs, wardrobes, dining tables, storage units) provide built-in supervision signal.
  • Provenance metadata links images back to product records that often include dimensions.

Note: for genuine second-hand / street photography conditions, consider combining ImageHub's studio/lifestyle images with real-world images from classified ad platforms (Gumtree UK, OLX) — which CrawlFeeds can source separately as a bulky waste image dataset. See custom data collection.

What to request

  • Categories: furniture (sofas, beds, chairs, wardrobes, dining tables), home appliances.
  • Filters: request lifestyle images specifically (not white-background only) — use the CLIP tags to filter for "lifestyle", "room", "interior".
  • Volume: 5,000-20,000 per category for classification, 50K+ for full recognition pipelines.
  • Custom: request Gumtree UK real-world bulky waste images separately via CrawlFeeds.

Key fields / metadata delivered

  • category_path: item type (sofas, beds, chairs, wardrobes)
  • clip_tags: style tags including lifestyle, interior, angle descriptors
  • site_name: source retailer (IKEA, Ashley Furniture, Wayfair, RH, Pottery Barn)
  • primary_image_flag: identifies main product view vs detail/lifestyle shots
  • image_dimensions: pixel dimensions for depth camera calibration reference

Furniture Dataset, Home and Kitchen Dataset, Ecommerce Image Dataset

Browse catalog Request custom dataset

Related use cases

Visual Search Engines
AI Product Tagging
Fashion Tech
Catalog Automation
Visual Intelligence & Image Optimization
Resale Item Recognition
All Use Cases

FAQ

Do you have images of furniture in real-room lifestyle settings rather than white backgrounds?

Yes — many retailer sources include lifestyle photography showing furniture in staged room settings. Request lifestyle-tagged bundles specifically for closer-to-real-world conditions.

Can I get dimension data alongside the images?

ImageHub provides image metadata but not product dimension specifications. For datasets with extracted dimension fields (length_cm, width_cm, height_cm), see CrawlFeeds custom data collection — we can source second-hand listings with seller-stated measurements.

Is this suitable for training depth camera calibration models?

Studio images provide useful reference for what items look like from known angles and distances. For depth camera ground-truth calibration, we recommend combining ImageHub imagery with real-world datasets from classified ad platforms where we can capture seller-stated dimensions.

What furniture categories are available?

Sofas, chairs, armchairs, beds, dining tables, coffee tables, wardrobes, storage units, outdoor furniture, and lighting. See the furniture dataset page for the full category breakdown and current image counts.

Can I get images from second-hand or classified ad platforms instead?

Yes — through CrawlFeeds custom data collection we can source images from Gumtree UK, OLX, and similar platforms showing items in real-world conditions. Contact us to scope a custom project.