![]() One of the first algorithms we are working on is actually Dead pixel detection. We are currently working on implementing some of our algorithms in GPU, which will give us more flexibility and more efficiency with some of our algorithms. Changing the sensitivity would require new analysis to be run. QScan can then only analyze dead pixel once, with a fixed sensitivy value. Also, as you probably know by now, QScan is storing each and every parameter that has been measured, in a DB, thus allowing the "single pass analysis" (analyze only once, and change your theshold settings as many times as you want). Dead pixel requires high processing power, and when working in CPU it would be difficult to make this sensitivity value selectable by the user. This sensitivity value is internal for now and it is fixed. ![]() There is a sensitivity value for the algorithm to identify pixels as candidates. This is just a basic explanation, but a lot more things are obviously examined to make a dead pixel a candidate. That is basically what you configure when you set the duration of the dead pixel detection If the same candidate appears in consecutive frames, it is given a "score" that increases for each consecutive frame where the same dead pixel is detected In every frame, QScan analyzes and flags "potential" dead pixels That is, same coordinate of dead pixels for all frames. ![]() As explained before, the current algorithm will only detect dead pixels that happen in several consecutive frames. We've been working very hard to get the latest release of QScan published (2023), and we've left the forum a bit unattended.Ībout dead pixel. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |