Image quality analysis for searches

摘要

Image analysis includes: determining, using one or more processors, an image quality score associated with an image, including: determining a foreground and a background in the image; calculating a set of one or more characteristic parameters of the image based on the determined foreground and background; calculating the image quality score based at least in part on the set of characteristic parameters, wherein calculating the image quality score comprises using an image quality computation model that has been pre-trained; and in response to a search query, generating a set of search results that includes a set the images, wherein inclusion of the images or ranking of the search results is based at least in part on image quality scores associated with the set of images.

主权项

1. A method comprising:
determining, using one or more processors, an image quality score associated with an image, including:
determining a foreground and a background in the image, comprising:
determining an outer frame zone of the image, wherein the outer frame zone is determined using a preset value from edges of the image;clustering pixels in the outer frame zone to obtain a plurality of color zones, wherein each color zone corresponds to a color type;determining seed zones;expanding the seed zones to obtain the background in the image; anddetermining the foreground;calculating a set of one or more characteristic parameters of the image based on the foreground and the background, wherein the characteristic parameters comprise a foreground-background contrast, and wherein the foreground-background contrast is determined by:
determining foreground color zones of the foreground and calculating an area ratio between an area of a foreground color zone and an area of the foreground;determining background color zones and calculating an area ratio between an area of a background color zone and the area of the foreground;calculating a Euclidean color-based distance between the foreground color zone and the background color zone; anddetermining the foreground-background contrast based at least in part on the Euclidean color-based distance and the area ratios; andcalculating the image quality score based at least in part on the set of characteristic parameters, wherein calculating the image quality score comprises using an image quality computation model that has been pre-trained; and in response to a search query, generating a set of search results that includes a set of images, wherein inclusion of the images or ranking of the search results is based at least in part on image quality scores associated with the set of images.