Good to know that OpenCV going to handle multiple tiff , but I still did understand imreadmulti documentation , if I have an image with unknown number of pages , who will direct me to use imread or imreadmulti , another issue imreadmulti read all pages in memory (verctor of MAT) , I have some books may reach to 1000 pages or more , loading all pages will consume most of device memory , I think it's better to be implemented as follow
1-Function to get number of pages .
2-If number of pages is one we can call imread
3-If Number of pages larger than 1 we can call imreadmulti
4-Do not load all pages in memory , just open the required index.
5-If the user need all pages , he will use the pages counter obtained from point 1 and loop all pages.