Boe technology group co., ltd. (20240242540). ACTION RECOGNITION METHOD AND DEVICE, MODEL TRAINING METHOD AND DEVICE, AND ELECTRONIC DEVICE simplified abstract

From WikiPatents
Jump to navigation Jump to search

ACTION RECOGNITION METHOD AND DEVICE, MODEL TRAINING METHOD AND DEVICE, AND ELECTRONIC DEVICE

Organization Name

boe technology group co., ltd.

Inventor(s)

Xianbin Liu of Beijing (CN)

Fanhao Kong of Beijing (CN)

Zhanfu An of Beijing (CN)

Zeyu Shangguan of Beijing (CN)

ACTION RECOGNITION METHOD AND DEVICE, MODEL TRAINING METHOD AND DEVICE, AND ELECTRONIC DEVICE - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240242540 titled 'ACTION RECOGNITION METHOD AND DEVICE, MODEL TRAINING METHOD AND DEVICE, AND ELECTRONIC DEVICE

    • Simplified Explanation:**

An action recognition method uses a pre-trained self-attention model to determine the probability distribution of a video being similar to various action categories, and identifies the target action category based on a preset threshold.

    • Key Features and Innovation:**
  • Electronic device samples image frames of a video for action recognition.
  • Utilizes a pre-trained self-attention model to determine probability distribution.
  • Identifies target action category based on preset threshold.
    • Potential Applications:**

This technology can be used in security systems, sports analysis, and video content recommendation algorithms.

    • Problems Solved:**

This technology addresses the need for accurate and efficient action recognition in videos.

    • Benefits:**
  • Improved accuracy in recognizing actions in videos.
  • Efficient processing of large amounts of video data.
  • Enhanced performance in various applications such as security and sports analysis.
    • Commercial Applications:**

The technology can be applied in security systems for surveillance, sports analysis for performance evaluation, and video content recommendation algorithms for personalized user experiences.

    • Prior Art:**

Prior research in the field of action recognition using machine learning models and attention mechanisms can provide insights into the development of this technology.

    • Frequently Updated Research:**

Stay updated on advancements in machine learning models for action recognition and attention mechanisms to enhance the performance of this technology.

    • Questions about Action Recognition:**

1. How does this technology improve the accuracy of action recognition in videos? 2. What are the potential commercial applications of this technology in various industries?


Original Abstract Submitted

an action recognition method includes: an electronic device sampling a plurality of image frames of a video to be recognized, and according to the plurality of image frames and a pre-trained self-attention model, determining a probability distribution of the video to be recognized being similar to a plurality of action categories; further, based on the probability distribution of the video to be recognized being similar to the plurality of action categories, the electronic device determining, from the plurality of action categories, an action category having a probability greater than or equal to a preset threshold as a target action category corresponding to the video to be recognized.