No self-introduction
Loading
Profile
Follow
codes (2)
Pyqt simple example - image browsing
no vote
Use Python to write a very simple picture viewer, interface is in Qt. Supported picture format jpg, TIF, bmp, gif. Main purpose is to make everyone what's written in Python, using Qt GUI.
crothen 0 0
1
BM3D image denoising
4.0
Three dimensional block matching (BM3D) algorithm: it first divides the image into blocks of a certain size, according to the similarity between image blocks, combines two-dimensional image blocks with similar structure to form a three-dimensional array, then processes these three-dimensional arrays by joint filtering method, and finally returns the processed results to the original image through inverse transformation, so as to get the denoised image Images. It not only has a high signal-to-noise ratio, but also has a good visual effect. Therefore, many denoising methods based on BM3D are proposed, such as: BM3D denoising based on wavelet transform, BM3D filtering based on Anscombe transform domain and so on. The general implementation process of this algorithm is as follows: for a noisy image, continuously take out the reference block, for each reference block, perform the following operations: Step 1. Basic estimation 1) block by block estimation. Each block I) in the noisy image is grouped. Find its similar blocks and aggregate them into a three-dimensional array. II) joint hard threshold. The three-dimensional array is transformed into a three-dimensional array, and the noise is reduced by hard thresholding of the coefficients in the transformation domain. Then the estimated values of all the image blocks in the group are obtained by inverse transformation, and then these estimated values are returned to their original positions. 2) Gather. For the overlapped block estimation, the weighted average is used to get the basic estimation of the real image. Step 2. Final estimation 1) block by block estimation. Each block I) in the basic estimation image is grouped. The positions of similar blocks in the basic estimation image are found by block matching. Two 3D arrays are obtained from these positions, one is from the noisy image, and the other is from the basic estimation image. II) joint Wiener filtering. The two 3D arrays are transformed into three dimensions, and the energy spectrum in the basic estimated image is used as the energy spectrum to filter the noisy 3D array. Then the inverse transformation is used to get the estimates of all the image blocks in the group, and these estimates are returned to their original positions. 2) Gather. For the overlapped local block estimation, the final estimation of the real image is obtained by weighted average.
crothen 0 0
1
No more~
Follows0
Fans0