Multi-Lens Live View Composition Selection With On-Device AI

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users face inconvenience and effort in manually merging image frames from multiple cameras, and the aesthetic quality of the merged images is often compromised.

Innovation Solution

An electronic device with multiple lenses and neural network models automatically selects and combines image frames based on composition preference information, storing them in different memories for live viewing and recording.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If users manually merge image frames acquired through multiple cameras, then the user has control over the merging process, but the user spends considerable effort and time to master image editing applications and the aesthetic quality is not ensured

Engineering Contradiction:
Improveaesthetic quality of merged imagesVSAvoiduser effort and time for manual merging
Core Design Contradiction:
Manufacturing precisionVSEase of operation

Solution Approach 1:

The system performs automatic image frame selection and merging without requiring user intervention. The electronic device autonomously evaluates composition quality using evaluation indices and automatically merges selected frames, eliminating the need for users to manually edit images while ensuring high aesthetic quality through algorithmic optimization

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical operations (user dragging, dropping, and adjusting image frames) with an automated computational system. The processor automatically evaluates composition quality using evaluation indices and performs merging operations, substituting human manual work with algorithmic processing that ensures consistent high-quality results

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Manufacturing precision

If the electronic device processes and stores multiple image frames from different lenses, then the image quality and composition preference are improved, but the memory usage and processing complexity increase

Engineering Contradiction:
Improveimage composition qualityVSAvoidprocessing and memory management complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent divides the memory management into distinct segments: first memory for storing original multiple image frames from different lenses, second memory for storing selected frames after evaluation, and third memory for storing final merged images. This segmentation organizes the complex data flow and reduces processing complexity by handling different stages separately

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary evaluation of image frames using composition evaluation indices before final merging. The processor pre-assesses each frame's composition quality and pre-selects optimal frames to be merged, reducing the complexity of the final merging operation by having already filtered and organized the input data

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4133715B1Electronic device and controlling method of electronic device
Publication Date: 2025.11.26 SAMSUNG ELECTRONICS CO LTD
  • EP4133715B1 patent drawingFigure 1
  • EP4133715B1 patent drawingFigure 2
  • EP4133715B1 patent drawingFigure 3~4C

AI summary

An electronic device and a controlling method of the electronic device are provided. The controlling method of an electronic device according to the disclosure includes the steps of, based on a first user input for or related to acquiring a live view image through a camera including a plurality of lenses different from one another being received, acquiring a plurality of image frames for each of the plurality of lenses and storing the image frames in a first memory, inputting the plurality of image frames for each lens stored in the first memory into a neural network model, by a predetermined time interval, and acquiring score information including composition preference information of each of the input image frames, selecting at least one lens among the plurality of lenses based on the score information, storing image frames acquired through the selected at least one lens in a second memory during the predetermined time interval, and based on a second user input for initiating recording of the live view image being received, storing an image related to the image frames stored in the second memory in a third memory until a time point when a third user input for ending the recording is received.