A method and device for detecting abnormal video stream events
A technology of abnormal events and detection methods, applied in the field of image processing, can solve problems such as inaccurate detection, high detection difficulty, and restriction of detection accuracy, and achieve the effect of improving accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0019] Please refer to figure 1 , the embodiment of the present application provides a method for detecting an abnormal video stream event, comprising the following steps:
[0020] 101. Input the training sample set into the deep learning neural network, and learn the trained deep learning neural network.
[0021] The deep learning neural network includes: multiple autoencoder machines stacked together, the training sample set is a collection of multiple training samples, and the training samples are extracted from training images.
[0022] It should be further pointed out that a large number of pictures or image blocks should be used to train the deep learning neural network to improve the judgment accuracy of the deep learning neural network. Generally speaking, when training a deep learning neural network, the images input into the neural network should be small image blocks with a preset size. Larger pictures or images are cut and divided into image blocks that can be us...
Embodiment 2
[0050] Please refer to Figure 6 , the embodiment of the present application provides a device for detecting abnormal video stream events, including:
[0051] The training stage input unit 30 is used for inputting the training sample set to the deep learning neural network, learning to obtain the model parameters of the deep learning neural network, and obtaining the trained deep learning neural network; wherein, the deep learning neural network includes: Multiple automatic encoder machines stacked together, the training sample set is a set of multiple training samples, and the training samples are extracted from training images.
[0052] The training phase learning unit 31 is configured to learn, according to the training samples, the shape information feature parameters, the motion information feature parameters, and the joint feature parameters of the shape information and the motion information of the training samples.
[0053] The discriminator construction unit 32 is us...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More