Search based on colors, shapes and textures

LTU technologies’ image similarity search enables automatic ranking of similar images. The similarity between images is based on the analysis of shapes, colors and textures.

Database images are ranked from most similar to least similar based on the quantitative distance between a query and the reference images.

A better way to search images

Similarity search is typically used in combination with categories and keyword search for large image databases such as online retail sites, stock photography portals and digital asset management systems. By combining textual search and categorization with similarity, browsing instantly becomes much more relevant and engaging.

• Find images that are similar in shape and color
• Bounce from image to image to expand search and visually browse through a catalog of images

Search on shapes and/or colors

Depending on the application, similarity search can be configured to perform searches based on:

- Shape and color combined equally - Shape and color with different percentage values - Exclusively shape or color only