A synthesis-based endoscopic video deblurring method
A deblurring and endoscopic technology, applied in the field of image processing, can solve problems such as unsatisfactory effects, and achieve the effects of convenient subsequent processing and improved reliability.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0035] Such as figure 1 As shown, in this embodiment, a method for deblurring endoscopic video based on synthesis is used to deblur the blurred frames in the number of video frames, and use a grid-based alignment algorithm to align the blurred frames in the number of video frames through the grid. The clear frame is divided into several rectangular blocks. Under the common constraints of the vertices of multiple rectangular blocks, the alignment of the blurred frame and the clear frame is realized by calculating the multiple homography matrices corresponding to the multiple rectangular blocks, and then the blurred frame is re-synthesized by the DPM algorithm. frames to make blurry frames clear. This scheme proposes a grid-based parameter-free motion estimation model, which improves the robustness of the gastroscope image alignment algorithm, and restores the blurred frame by using the clear frame adjacent to the blurred frame, which is conducive to finding the most similar ima...
Embodiment 2
[0037] On the basis of the foregoing embodiments, in this embodiment, a method for deblurring endoscopic video based on synthesis includes the following steps:
[0038] Step S1: Input the endoscopic video t(x, y), the endoscopic video t(x, y) contains several video frames, assuming that the Nth frame is the fuzzy frame to be processed, select the fuzzy frame to be processed f (x, y), in frames N-10 to N+10, select a frame adjacent to the fuzzy frame f(x, y) with the best image quality as the clear frame g(x, y).
[0039] Step S2: Initialize parameters.
[0040] Step S3: Align the sharp frame g(x, y) to the blurred frame f(x, y) using a grid-based alignment algorithm. Using the grid-based alignment algorithm, the non-rigid inner surface of the human stomach can be better simulated, and the distortion of the image can be effectively reduced, thereby improving the image quality after blurred frame restoration. It includes the following steps:
[0041] Step S31: if figure 2 A...
Embodiment 3
[0060] Such as figure 2 As shown, on the basis of the above embodiment, in this embodiment, in the step S31, when estimating the motion vector, it is judged whether the image in the region belongs to the texture-rich type, and if so, the Surf image feature points are selected to estimate the motion vector, such as figure 2 As shown in (c), . Conversely, if the image in this area is not rich in texture but a homogeneous area, use the optical flow point to estimate the motion vector, such as figure 2 (b) shown.
[0061] In this embodiment, the feature points of the Surf image are extracted by using the Surf feature point extraction algorithm. The Surf feature point extraction algorithm is a very famous and classic feature point extraction method in the field of image processing and computer vision. Surf is the abbreviation of Speeded Up Robust Features. Proposed by Hebert Bay in 2008. In the implementation of this algorithm, the algorithm is directly called to extract th...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


