Browse Prior Art Database

System Method and a Learning Engine for Optimum Photo Selection Disclosure Number: IPCOM000248082D
Publication Date: 2016-Oct-24
Document File: 3 page(s) / 248K

Publishing Venue

The Prior Art Database


Disclosed are a system and method that utilize a data analysis model, a learning engine, and an associated device application to indicate to users in real time whether a photograph is satisfactory based on the user’s preferences.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

System Method and a Learning Engine for Optimum Photo Selection

In an effort to capture a good image of a subject, event, location, etc. a user often takes multiple photographs (photos) using a mobile device (e.g., smart phone, tablet, etc.). The problem is that the user does not or cannot immediately view the picture; therefore, the user cannot determine whether the quality of the image as a whole or some attributes of the image are satisfactory or unsatisfactory.

The novel solution uses the concept of machine learning to identify the photographs that best satisfy the user's preferences. The core concept is to implement an application (app) that autonomously learns the user's preferences for photographs (i.e.,

what the user does and does not like). The solution applies an algorithm to learn which photographic attributes to find and determine as desirable relative to the target audience. The app then indicates to the user whether the captured photo is a "good" or "bad" image.

As input to the machine learning, the user uploads to the algorithm photographs that represent both desirable and undesirable attributes. When the user captures an image using any device, the app recognizes the attributes as desirable or undesirable, and then alerts the user as to whether the image is satisfactory or requires the user to take action to improve the quality or appearance of the photo.

This application allows the user to know whether the photo is satisfactory, without

actually seeing it. This system and method can identify certain positional outlines of the user, the lighting the user generally wants in a photo, the ratio of the user (if in the photo) to the backdrop, and general placement of other people in the photo. The application, in real time, indicates to the user the likelihood of the user appreciating the photograph. For example, the device can flash a light with one color to indicate a good photograph and another color to indicate a bad photograph. If the photo has a low likelihood of satisfying the user, then the app allows the user the opportunity to make changes (e.g., positions, lighting, etc.) to improve the quality of the photo. If the user actually likes a photograph that the app deemed poor, then the algorithm uses those photographic attributes to learn from the user and change the level of appreciation of the photo. This helps the algorithm meet the user's needs and expectations and improve assessment over time.

The novel system and method utilize a data analysis model. The steps for implementation f...