Face Image Matrix Navigation for Video Content Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulty in recognizing the content of video content data due to the lack of effective methods for efficiently selecting the reproduction start position, especially with long video files, as title names alone are insufficient and fast-forwarding is time-consuming.
Innovation Solution
An electronic apparatus extracts face images from video content data, displays them in a matrix format with timestamp information, and adjusts cutout ranges to prevent protrusion, allowing for efficient navigation and selection of reproduction start positions based on face images and attribute sections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If the cutout range of the face area is decided to make the positions and sizes of faces constant, then the uniformity of face images is improved, but the cutout area protrudes outside the frame when the face is at the end part
Solution Approach 1:
The patent applies local quality by making different parts of the face extraction process have different properties. When a face is detected at the end part of the frame, the cutout range is adjusted locally to fit within the frame boundaries, while faces in the middle of the frame maintain the standard constant size and position. This allows the system to maintain uniformity where possible while adapting locally to prevent protrusion.
Solution Approach 2:
The patent introduces dynamics by making the cutout range adjustable rather than fixed. The system dynamically determines the cutout range based on the face position within the frame. When a face is detected at the end part, the cutout range is modified to prevent protrusion, while maintaining constant size and position for faces in the middle, thus creating a dynamic adaptation to different spatial contexts.
2Quantity of substance
If the total time length of video content data is long, then the content information is comprehensive, but the time required to reproduce and navigate the content increases
Solution Approach 1:
The patent applies segmentation by dividing the long video content into smaller manageable units represented by face images. Each face image serves as a segment marker that users can quickly identify and select. This segmentation allows users to navigate through comprehensive content without having to replay entire long sections, significantly reducing navigation time while preserving access to all content information.
Solution Approach 2:
The patent implements preliminary action by pre-extracting and displaying face images from the video content before user selection. These face images are prepared in advance as navigation markers, allowing users to immediately jump to desired sections without time-consuming fast-forwarding. The preliminary extraction of face images creates a ready-to-use index that speeds up content access.
Data Source
AI summary
According to one embodiment, an electronic apparatus extracts face images of persons from video content data and outputs timestamp information indicating time points at which each extracted face image appears in the video content data, and displays face images in each column of a plurality of face image display areas arranged in a matrix based on the time stamp information. The apparatus detects presence or absence of a face area in each frame consisting of the video content data and decides a cutout range of the detected face area. And, the apparatus adjusts a case in which the cutout range of the decided face area protrudes outside the frame.