Object-Tracked Video Summaries for Lower Bandwidth Delivery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video processing systems are inefficient in extracting and summarizing relevant content related to specific object types, requiring significant manual effort and resources, and result in high bandwidth and memory usage for transmitting full videos.
Innovation Solution
A video generation system that processes input videos to extract sub-videos showing target objects, generates tracking data to identify and track these objects, and combines them into topical videos using a video template, reducing the duration and size of the output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If full videos are transmitted for content delivery, then complete video content is available, but bandwidth and memory usage increase significantly
Solution Approach 1:
The system extracts only the necessary portions of video content related to specific target objects rather than transmitting complete videos. Video processing systems identify and extract sub-videos containing target objects of interest, removing irrelevant portions before transmission, thus reducing bandwidth and memory requirements while preserving essential information.
Solution Approach 2:
The patent segments video content into multiple sub-videos based on target object identification. Each sub-video contains content related to specific target objects, allowing selective transmission and processing of only relevant segments rather than handling complete video files, thereby reducing resource consumption.
2Manufacturing precision
If manual video processing is performed to extract relevant content, then precise object extraction is achieved, but significant manual effort and resources are required
Solution Approach 1:
The system implements automated video processing that performs object identification, tracking, and sub-video extraction without manual intervention. The video processing system autonomously analyzes video content, identifies target objects using tracking data, and generates condensed videos, eliminating the need for manual processing while maintaining high extraction accuracy.
Solution Approach 2:
The patent replaces manual mechanical processing with automated computational systems. Machine learning models and computer vision algorithms substitute human operators, automatically performing object detection, tracking, and video segmentation tasks with high precision and improved processing efficiency.
3Loss of energy
If video duration is reduced to save bandwidth, then transmission efficiency improves, but video content completeness may be compromised
Solution Approach 1:
The system applies different quality levels to different portions of video content based on relevance. Sub-videos containing target objects of interest are preserved with high quality, while irrelevant portions are removed. This selective quality approach ensures that bandwidth is consumed efficiently while maintaining completeness of important content.
Solution Approach 2:
The video processing system performs preliminary analysis to identify and extract relevant content before transmission. By pre-processing videos to identify target objects and generate tracking data, the system prepares condensed sub-videos in advance, ensuring that only necessary content is transmitted with full quality, avoiding post-transmission quality compromises.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating videos. In one aspect, a method comprises: receiving: (i) an input video comprising a sequence of video frames, and (ii) data indicating a target object type; processing the input video to generate tracking data that identifies and tracks visual locations of one or more instances of target objects of the target object type in the input video; generating a plurality of sub-videos based on the input video and the tracking data, including: for each sub-video, generating a respective sequence of sub-video frames that are each extracted from a respective video frame of the input video to include a respective instance of a given target object from among the identified target objects of the target object type; and generating an output video that comprises the plurality of sub-videos.