Unlock instant, AI-driven research and patent intelligence for your innovation.

Unsupervised monitoring video prediction frame anomaly detection method

A technology for monitoring video and anomaly detection, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as performance differences, and achieve the effects of low latency, fast speed, and reduced blurring.

Active Publication Date: 2022-02-18
XI AN JIAOTONG UNIV
View PDF7 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Hand-designed features usually have strong interpretability or certain physical meaning, but there are significant performance differences compared with deep learning features based on representation learning

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Unsupervised monitoring video prediction frame anomaly detection method
  • Unsupervised monitoring video prediction frame anomaly detection method
  • Unsupervised monitoring video prediction frame anomaly detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056] The present invention will be described in detail below in conjunction with the accompanying drawings.

[0057] The present invention utilizes a unified generating confrontation network (including a generator and two discriminators) to accurately predict video frames, utilizes the limitation of cyclic retrospectiveness to keep the consistency of predicted past frames and future frames with video sequences, and reduces Predict how blurry the frame will appear. And an Attention Weight Map is proposed to alleviate the foreground-background imbalance problem in anomaly detection.

[0058] For prediction-based video anomaly detection methods, it is usually assumed that there is some regular contextual connection in a continuous normal video, and this dependency can be learned and future frames can be better predicted. In contrast, a continuous anomalous video often violates these dependencies, making future frames unpredictable. Therefore, prediction errors for future vide...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an unsupervised monitoring video prediction frame anomaly detection method. The method comprises the steps: calculating a PSNR index through a prediction frame generated by an anomaly detector and a real frame, performing normalization to obtain a score of a video frame, if the score of the video frame is larger than a set threshold value, judging that the video frame is abnormal, and otherwise, judging that the video frame is normal. As the anomaly detection judgment mode is to calculate the PSNR value, and whether the target frame is abnormal or not can be judged only by inputting a plurality of frames, the speed is high, and the delay is low, and the anomaly judgment capability depends on the modeling capability of a generator on the normal video frame by modeling the distribution of the normal video frame and taking the abnormal video frame as outlier detection, so that the false alarm rate of detection is low.

Description

technical field [0001] The invention belongs to the technical field of video detection, and in particular relates to an unsupervised surveillance video prediction frame abnormality detection method. Background technique [0002] In the face of all kinds of security threats and emergencies that are constantly emerging in daily life, the use of video surveillance as a tool for security has highlighted its strong advantages. In recent years, with the rapid development of social economy and the continuous popularization of video sensing technology, surveillance systems have been widely used in various public places such as public security, subways, communities, and campuses. However, the massive video data generated by the rapidly growing video surveillance system has brought great challenges to video anomaly event detection based on human interpretation. The traditional way of relying on manual viewing of post-event monitoring video records to find abnormalities not only consu...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06V20/40G06N3/04G06N3/08
CPCG06N3/049G06N3/088G06N3/044G06N3/045
Inventor 李刚李慧斌张凡何平
Owner XI AN JIAOTONG UNIV