Journal: Photonics

Article Title: Active Polarization Imaging for Cross-Linear Image Histogram Equalization and Noise Suppression in Highly Turbid Water

doi: 10.3390/photonics10020145

Figure Lengend Snippet: Figure 3. The flowchart of underwater polarization differential imaging enhancement based on CLAHE cross-linear image histogram equalization and joint noise suppression.

Article Snippet: With the known results given above, Equation (5) can be used to calculate the N processed target signal light results obtained from Ipromin,i(i = 1, 2, . . . , N), as follows: Si(i = 1, 2, . . . , N) = 1poptimumscat −p optimum targ [ Ipromin,i(i = 1, 2, . . . , N) ( 1 + poptimumscat ) −Ipromax ( 1− poptimumscat )] (15) The N Si(i = 1, 2, . . . , N) obtained by using Equation (15) are then used with CLAHE to perform histogram equalization; so, there are, Sproi (i = 1, 2, . . . , N) = CLAHE[Si(i = 1, 2, . . . , N)] (16) Photonics 2023, 10, 145 8 of 19 Next, in order to reduce the gained noise in the target signal light, the N histogram equalized Sproi (i = 1, 2, . . . , N) is subjected to multi-frame averaging processing, and then, bilateral filtering is used, BF ( Sproaverage ) = BF ( N ∑ i=1 Sproi / N ) (17) Here, BF(·) denotes the bilateral filtering operation, and BF ( Sproaverage ) is the result of polarization enhancement by equalizing the histogram of the cross-linear polarization image and combining the noise suppression processing.

Techniques: Imaging

Journal: Photonics

Article Title: Active Polarization Imaging for Cross-Linear Image Histogram Equalization and Noise Suppression in Highly Turbid Water

doi: 10.3390/photonics10020145

Figure Lengend Snippet: Figure 5. (a) is the raw intensity image Imin; (b) is the result of processing (a) using Li’s method [25]; (c) is the result of processing (a) using CLAHE [11]; (d) is the normalized grayscale histogram of (a); (e) is the normalized grayscale histogram of (b); (f) is the normalized grayscale histogram of (c); (g) is the grayscale value row indexes of the extreme points in (b); (h) is (b) the grayscale normalization result after removing the extreme point; (i) is the normalized grayscale histogram of (h).

Article Snippet: With the known results given above, Equation (5) can be used to calculate the N processed target signal light results obtained from Ipromin,i(i = 1, 2, . . . , N), as follows: Si(i = 1, 2, . . . , N) = 1poptimumscat −p optimum targ [ Ipromin,i(i = 1, 2, . . . , N) ( 1 + poptimumscat ) −Ipromax ( 1− poptimumscat )] (15) The N Si(i = 1, 2, . . . , N) obtained by using Equation (15) are then used with CLAHE to perform histogram equalization; so, there are, Sproi (i = 1, 2, . . . , N) = CLAHE[Si(i = 1, 2, . . . , N)] (16) Photonics 2023, 10, 145 8 of 19 Next, in order to reduce the gained noise in the target signal light, the N histogram equalized Sproi (i = 1, 2, . . . , N) is subjected to multi-frame averaging processing, and then, bilateral filtering is used, BF ( Sproaverage ) = BF ( N ∑ i=1 Sproi / N ) (17) Here, BF(·) denotes the bilateral filtering operation, and BF ( Sproaverage ) is the result of polarization enhancement by equalizing the histogram of the cross-linear polarization image and combining the noise suppression processing.

Techniques:

Journal: Scientific Reports

Article Title: 3D histopathology of human tumours by fast clearing and ultramicroscopy

doi: 10.1038/s41598-020-71737-w

Figure Lengend Snippet: Computational improvement of cancer sample recordings. ( a ) Image processing chain for cancer biopsies. The UM image stacks are contrast enhanced by contrast limited histogram equilibration (CLAHE), followed by stripe artifact removal using a matched 2D Fourier transform slope filter and unsharp masking. ( b1 ) Representative slice of an UM data set obtained from a breast cancer biopsy before post-processing. ( b2 ) Same as in ( b1 ) after contrast enhancement using CLAHE. The visibility of information encoded in small brightness differences is clearly enhanced. ( b3 ) Stripe artifacts generated during UM recording have been removed via a matched 2D Fourier domain slope filter. ( b4 ) Finally, the image is slightly sharpened via unsharp-masking to further enhance the visibility of fine details. ( c ) UM recordings often exhibit stripe shaped artefacts originating from light absorbing structures that are persistent to the clearing procedure. By obstructing the light sheet these structures produce visible shadows that can include an angle α with the horizontal image edges depending on the camera orientation. ( d ) To remove the stripe artefact the images are Fourier transformed and multiplied with a filter mask cutting out a pie-slice shaped piece of the spectrum matching the angular direction α of the stipes. After inverse transformation and rescaling a stripe suppressed image is obtained. ( e1 ) Design of the pie shaped filter. The angular direction α of the stripes corresponds to an angle of 90-α in the 2D power spectrum. The angular direction and the shape of the pie slice filter can be optimized in the software by varying α and the distances d 1 , d 2 , w 1 , and w 2 . This allows to match bandwidth and direction sensitivity of the filter in order to find a parameter combination providing best possible stripe suppression at minimal costs of blurring artefacts or ringing. ( e2 ) To reduce ringing artefacts due to a hard frequency cutoffs, the edges of the pie shaped filter exhibit a smooth Gaussian transition profile. All image processing steps were performed using custom-made software written in MATLAB (MathWorks, Germany) and Visual Basic.Net (Microsoft, USA). The programs can be obtained from klaus.becker@twien.ac.at upon reasonable request.

Article Snippet: The images were contrast enhanced using the CLAHE algorithm provided by the MATLAB Image Processing Toolbox (MathWorks, Germany).

Techniques: Generated, Transformation Assay, Software