Progressive single image rain and snow removal method, device and readable medium
CN118864287BActive Publication Date: 2024-11-29HUAQIAO UNIVERSITY
Patent Information
- Application Number
- CN202411319738.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-09-23
Smart Images

Figure CN118864287B_ABST
Abstract
The present invention discloses a progressive single-image rain and snow removal method, device and readable medium, which relates to the field of image processing, including: constructing and training an image rain and snow removal model based on a recurrent dense network to obtain a trained image rain and snow removal model, wherein the image rain and snow removal model includes a plurality of stacked internal and external densely connected subnetworks, each of which includes a plurality of stacked internal and external densely connected blocks, each of which includes a long short-term memory module, a first convolutional layer, a channel attention module and a second convolutional layer connected in sequence; inputting a degraded image into the trained image rain and snow removal model, the restored estimate outputted by the current stage of the internal and external densely connected subnetwork and the degraded image are inputted into the next stage of the internal and external densely connected subnetwork, and the last stage of the internal and external densely connected subnetwork outputs a restored image. The present invention solves the problem of artifacts existing in images restored by the current progressive rain and snow removal method.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
Image defogging and rain-removing algorithm based on selective attention mechanism
CN111652812A
Image rain removing method based on airspace group enhancement mechanism and long-short-term memory network
CN111861926A
