| 摘要 |
Systems and methods are presented for identifying unwanted photos and videos on a device. Such systems may contain various modules to do one or more of: (i) review the image quality of photos and videos, (ii) review a user's interaction with photos and videos, (iii) review the properties of photos and videos, (iv) review the user's profile/preferences as to photos, or (v) review information obtained from additional sources regarding the user, his or her photos, and any interaction therewith. With one tap, an exemplary application can identify all of a user's unwanted photos, such as, for example: photos with bad lighting, blurry shots, similar photos, and even photos that the system predicts a user may find boring. Any photos that an exemplary system is unsure about may be left for the user to review and act upon. By analyzing user decisions exemplary systems become progressively more accurate and thus improve their predictions over time. |
