Search of duplicates is performed in two stages. After you set the source and output folders, the program will perform a preliminary analysis of the photos. This process is performed only once and does not depend on the value of similarity. Speed image processing is about 10MB/sec.
The actual process of finding duplicates starts after you press Start processing button. The program sorts the duplicates by groups. In each group the most similar photos are placed. It is guaranteed that the leftmost picture is different from other photos no more, than you set in the similarity parameter. It should be noted that the similarity does not match pixel-area coincidence. It does not match the color palette. This is some integral parameter. Therefore, photos with different brightness, scale, orientation or even a black and white and color photos can be considered as similar photo. The program accepts rotated images on 90 and 270 degrees as equal. The comparison result is weakly dependent on the quality of the image. Black and white copies of color photographs give a difference of 1-2 percent. When the value is less than 90 percent program selects a similar picture, but not a duplicates, as you it understand. We recommend starting with the value of the parameter 98 percent. Window is shown in the figure below. The program allows you to compare photos of different formats. Window is shown in the figure below.
For a detailed analysis, you can use the histogram. You can compare the photos or their histograms. For each photo, you can specify the action to be performed with this photo. All actions are performed only after pressing the Start processing. The figure below shows the window for selecting actions to the image.