Browse Prior Art Database

AUTOMATIC RECOMMENDATIONS OF IMAGES AND VIDEOS TO DELETE

IP.com Disclosure Number: IPCOM000239057D
Publication Date: 2014-Oct-06
Document File: 9 page(s) / 219K

Publishing Venue

The IP.com Prior Art Database

Abstract

Various implementations provide a system automatically providing recommendations of images, such as photos and videos, to delete from a user's collection. In various examples, the system can examine characteristics of images and score the images. Based on the scores, the system presents a subset of the images to a user as recommendations for deletion. Videos can also be recommended for deletion based their frames being examined as images. Such features allow a large collection of images and videos to be better managed for a user by automatically pointing out images and videos that are less desirable to keep.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 16% of the total text.

Page 01 of 9

AUTOMATIC RECOMMENDATIONS OF IMAGES AND VIDEOS TO DELETE

ABSTRACT

Various implementations provide a system automatically providing recommendations of images, such as photos and videos, to delete from a user's collection. In various examples, the system can examine characteristics of images and score the images. Based on the scores, the system presents a subset of the images to a user as recommendations for deletion. Videos can also be recommended for deletion based their frames being examined as images. Such features allow a large collection of images and videos to be better managed for a user by automatically pointing out images and videos that are less desirable to keep.

DESCRIPTION

    The present disclosure describes features related to automatically providing recommendations of images and videos to delete.

Analyzing and Scoring Images

    A system can examine one or more collections of images, such as photos stored in albums on client devices (phone, tablet, desktop, etc.) and/or photos stored on Internet servers such as a photo collection service, social networking service, etc. For example, the system can examine all the photos and videos owned or accessible by a user. The system determines particular characteristics of the examined images, which can include technical quality characteristics of images, content of images, and similarity of images.

    The technical quality of images can be described by characteristics such as an amount of blur in an image, amount of color noise in an image, and the exposure level of an image that are determined by the system using image analysis techniques. Some implementations can determine other types of characteristics such as image size, resolution, color cast, etc. The content of images can include characteristics such as recognized types of objects, including faces, human figures, landscape features, animals, or other types. Depicted content in the image can be determined using object recognition techniques (e.g., facial recognition) and similar techniques. The similarity of images can include the amount of similarity of each image to any

1


Page 02 of 9

other images in the examined collection, such as similarity of depicted content, and/or similarity of technical characteristics (color cast, blurriness, exposure, noise, etc.). The system can compare pixels, content and regions in images to find such similarities.

    In some examples, the system can assign an overall score to each image based on the examined characteristics of the image. A score indicates the overall desirability of keeping or retaining the image, and can include indications of a visual quality of an image, image content desirability, and/or a visual similarity of an image to other images in the collection. In the examples described, a higher score indicates higher quality and desirability of the image, and a lower score indicates the opposite.

    For example, the system can determine the score based at least in part on visual quality c...