Video Search Interface for Continuous Playback Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video search methods require users to repeatedly navigate back and forth between video playing pages and search result pages, reducing efficiency and continuity of video consumption.
Innovation Solution
A video search method that allows seamless switching between videos within the search results interface, enabling continuous playback without exiting the current display, using a first area for video playback and a second area for displaying search results, with options for manual or automatic video transitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional video search methods are used, then the system structure is simple, but the search accuracy is low and user experience is poor
Solution Approach 1:
The patent segments the video search system into multiple independent modules: video preprocessing module, feature extraction module (with audio, visual, text separate streams), similarity calculation module, and ranking module. This segmentation allows each module to specialize in specific tasks, improving overall search accuracy while maintaining manageable system complexity through modular design.
Solution Approach 2:
The patent introduces multiple intermediary components: audio features as intermediaries between video content and search queries, visual features as intermediaries for content understanding, and text features as intermediaries for semantic matching. These intermediaries transform raw video data into structured representations that enable accurate similarity calculation and ranking.
2Measurement precision
If multiple feature extraction methods are applied, then search precision is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-extracting and storing audio features, visual features, and text features during video preprocessing, rather than extracting them in real-time during search. The system pre-computes similarity scores and ranks videos before user queries arrive, significantly reducing processing time during actual search operations while maintaining high precision through comprehensive feature analysis.
Solution Approach 2:
The patent implements periodic action by updating video features and similarity rankings at scheduled intervals or when new videos are added, rather than continuously processing all videos for every query. This periodic updates approach maintains search precision while avoiding unnecessary continuous processing that would waste time and resources.
3Ease of operation
If comprehensive video analysis is performed, then user experience is improved, but computational resources are excessively consumed
Solution Approach 1:
The patent applies partial action by selectively extracting only the most relevant features (audio, visual, text) needed for video search, rather than analyzing all possible video attributes. The system performs sufficient feature extraction to achieve good user experience while avoiding excessive analysis of irrelevant video properties, thus optimizing computational resource consumption.
Solution Approach 2:
The patent changes parameters by transforming raw video data into multiple feature representations (audio spectra, visual descriptors, text embeddings) with different characteristics. This parameter transformation allows the system to analyze videos comprehensively for good user experience while working with optimized feature representations that reduce computational burden compared to processing raw video data directly.
Data Source
Figure 1A~1B
Figure 2A~2B
Figure 3~4
AI summary
The present disclosure provides a video search method and apparatus, a device, a storage medium and a product program. The method comprises: acquiring a first search result about video search information, wherein the first search result comprises a plurality of videos; playing a first video in the first search result in a first area of a first display interface, and displaying the first search result in a second area displayed on the first interface; and in response to a first trigger operation for the first area, playing a second video in the first search result in the first area.