E-Song Multimodal Search Engine
A high-performance AI search engine designed for professional search purposes.
What Is E-Song Multimodal Search Engine?
The E-Song Multimodal Search Engine is a high-performance AI-powered tool designed for professional search applications. It incorporates advanced AI capabilities, including content understanding, automatic content collection, and automatic clue organization. This search engine can effectively search for specific objects, items, faces, text, voice, and event content. The search sources can include images, videos, social media, and more.
Notably, the E-Song Multimodal Search Engine stands out for its exceptional search speed and content comprehension. It can handle up to 100 million face searches per second. Additionally, it can interpret image content, such as inferring how a person might have looked at a younger age or recognizing their illustrated depiction. The engine can also perform event searches by analyzing event details, retrieving related information from media reports, and automatically generating reports.
Highly
Intelligent
Capable of
Web-Wide Search
Image Content
Understanding
Auto-Content
Collection
Auto-Clue
Organization
Auto-Report
Generation
Applicable Situations:
Fast and
Large-Scale Searches
Searches with Incomplete
Object Information
Event Clue
Collection
Applicable Situation 1
In a large surveillance area, quickly search for objects in real-time monitoring data or historical surveillance data. For example: quickly locating a lost child in an amusement park, or police searching for a suspect in a city-wide public camera system.
The E-Song Multimodal Search Engine can be deployed on edge boxes or servers. When used with an AI Camera, it can perform up to 100 million face searches per second.
Usage scenarios:
Real-time people search/tracking
Rapid evidence collection
Top recognition accuracy
Higher than well-known companies
E-Song：76.1%
Famous AI manufacturers：59.4%
（Conclusion: Not the same person）
Identifying extreme situations
Accurate recognition at extreme angles/high occlusion
3D head reconstruction and recognition
Mask recognition
Aging recognition
Recognition of people's appearance in different periods
High fault tolerance
Maintain high-precision recognition under different cameras/image quality
Dark Images
Phone
Mobile Camera
Applicable Situation 2
Applicable in situations where the information is incomplete or the content is complex. Tasks can be assigned in a conversational format, similar to ChatGPT, and corresponding results can be provided based on the conversation's needs.
The powerful AI capabilities can understand the content of images, videos, or documents in the database, enabling the AI to recognize events. It can search for related images, videos, written records, and other content.
Usage scenarios:
Search for information in the database
Large-scale search/tracking
Evidence/clue collection
Search multiple objects
People, Objects, Buildings, Text, Audio, Event
ChatGPT-style search
Conversational questions, automatic content organization
Image content understanding
Search for various targets
Applicable Situation 3
AI can continuously read and search public social media content 24/7. It can track clues, automatically organize the search results, and generate log reports.
Generates activities description and routines
from search results
Usage scenarios:
Evidence/clue collection
Historical information collection
Search all Internet information
Collect the latest information according to conditions
Auto-generation of summary /detailed reports
Generate reports in various formats
Image content understanding /judgment
Understand the relationship between picture details and content and judge rationality
Inappropriate content found
Event-related image
