Emotion recognition AI enables the analysis of content and marketing responses on mobile
A technology has been developed that utilizes emotion recognition artificial intelligence (AI) to analyze responses to content and marketing, including advertisements in mobile applications.
Semicon Networks (CEO Hong Kyeong Yi) has developed 'PEPERONE,' a technology based on AIMDVF (AI-based Mood Detection through Voice and Face) v1.0, which identifies a speaker's emotions in real time through voice and facial expressions. It can be utilized for the development of mobile-based services.
Emotion Recognition AI PEPERONE: Example Analysis of 8 Emotion Keywords
Facial expression analysis is enhanced by AI training using 360,000 facial images obtained from both public datasets provided by the Massachusetts Institute of Technology (MIT) and proprietary data collection efforts. The AI has been trained to recognize eight primary emotions, including joy, sadness, and anger, improving its analytical capabilities. Voice recognition is also trained using MIT datasets along with additional collected data.
PEPERONE supports both Korean and English and is designed to recognize facial expressions of both Asian and Western users.
The system enhances analysis accuracy by structuring prediction values derived from four neural networks that analyze voice and facial expressions in a multi-layered architecture. It is used as a tool to monitor consumer responses to content such as news articles and advertisements on mobile devices. By analyzing facial expressions while users engage with mobile screens and ambient voice data, the AI determines how effectively the content resonates with users. This enables businesses to assess the effectiveness of their content and marketing strategies.
The technology can be integrated into AI speakers and robots to recognize the speaker's emotional state and facilitate more engaging and intimate conversations. AI robots equipped with PEPERONE can detect the direction of the speaker's voice, establish eye contact, and enhance user engagement, making interactions more natural and personal.This technology is also applicable to companion robots and similar devices.
Semicon Networks recently applied for 14 patents related to PEPERONE's emotion recognition technology, including a method for analyzing consumer emotions to evaluate satisfaction with displayed content.
PEPERONE's core technology is now in the commercialization state, and the company is actively targeting domestic and international mobile app developers, as well as AI speaker and robot manufacturers.
Baek Dae Won, Vice President of Semicon Networks, stated, "PEPERONE is a key technology for detecting user emotions through facial expressions and voice. We plan to supply it to companies developing mobile applications for portals and e-commerce, as well as manufacturers of AI speakers and robots."
댓글
댓글 쓰기