The general approach to sharing data between OpenVX* and media/graphics APIs like Intel® Media Server Studio or OpenGL* is based on sharing the system memory. That is, in your code you should map or copy the data from the API to the CPU address space first. Then the resulting pointer can be wrapped as OpenVX image. Alternatively, the data (the pointer references to) can be copied to OpenVX object, to unlock the original data (and return ownership to another API).
Similarly, OpenVX data can be accessed on the host and its bits copied or wrapped to the data object of another API. Notice, though, that most APIs do not support memory layouts with strides between pixels that are possible in OpenVX, so your data sharing routine should respect the strides.