The image text1.png after replacing the yellow background with a white one.

The image t2.png after thresholding it at level 80 in order to make the background white while keeping the text legible.

The image t3.png after thresholding it at level 56 in order to make the background white. Text legibility had to be sacrificed as the gradient was much stronger on this image compared to t2.

The image barbotan.jpg at threshold level 120. Although the image doesn't have the background fully removed, decreasing the threshold any further makes the text less legible and I find this to be a good balance. Decreasing it past 95 makes it nearly illegible.

The image lines.png after running a Sobel Feldman edge detection algorithm using the website Pinetools. The black and grey rectangles are replaced with white space with black borders around where they used to be to signify their edges. Also the lines on making up the 'X' have disappeared leaving an outline around its edge instead similar to what happened to the rectangles.

The image towera.png under different noise reduction methods. All methods are built into Photoshop. The dust and scratches filter (with radius set to 1) does the best job of removing the noise while preserving image clarity in this instance. However, it does so at the cost of making the image more blurry overall. A balance can be achieved between the threshold values of 30 and 55. The despeckle filter does the best at preserving image detail but performs poorly in terms of actual noise reduction and merely tries to blend the noise with the background. However, the median filter (with radius set to 1) fares the worst out of all of them by removing most of the image detail in an attempt to remove the noise. Although it is successful in removing the noise, a similar effect with better image clarity can be achieved using the dust and scratches filter.