Just think about how you "see" an object.
Either by reflect or by self generation an object emits light from every point on its surface.
These rays of light emanate from a point and so are diverging.
Some of these rays of light pass through the cornea and lens system and hit the retina.
For a "perfect" image to be formed all the rays of light from one point on the object which enter the eye must hit one point on the retina and they all must have taken the same time to travel from the object to the retina (or arrive in phase).
Also the light from the points on the object which are nearest neighbours to the point under consideration on the object must arrive at the retina at points which are in the same relative positions to the point under consideration as they were on the object.
In the case of the reflection in a mirror the eye can only process the reflected light which "appears" (back produce the rays entering the eye) to come from a region behind the mirror which we call a virtual image of the object in front of the mirror. The virtual image acts like a light emitting object as far as the eye is concerned. Thus, virtual images are essentially seen due to the action of the eye.
Again there is a one to one correspondence between points on the object in front of the mirror and points on the virtual object behind the mirror.
The light which leaves the object can be catagorised in terms of the direction it is moving, its wavelength/frequency/colour and its intensity.
An optical system can be "designed" to bring that light together to reproduce an entity, the image, which is a point to point reproduction of the object.
So if the retina was not there the light rays which had formed the image on the retina would carry on moving in a divergent manner and those light rays would appear to come from where the image had been formed on the retina.