A standard rule in 3D rendering is that if two users are at two different positions in space, they will have two different renderings. However, a complete model of the space must be known to respect this rule and create different renderings. With a single 360° image, only the view of the space at the camera position is available. This restriction is challenging when creating a collaborative environment with 3D avatars from a 360° image. Here, we propose a method to integrate avatars in a coherent way in a 360° image. To add interaction between users, a laser pointer metaphor is developed to select elements on the 360° image.