Conclusion

This concludes our description of the IMAGINE I vision system. Starting from surface depth and orientation information for the visible surfaces in the scene, it could:

• produce an identity-independent segmentation of the objects in the scene,
• describe their three dimensional properties,
• select models to explain the image data,
• methodically pair the data to model features (while extracting the object's spatial position and explaining missing features arising from occlusion or object position) and
• verify the existence and identity of the instantiated hypotheses

for non-polyhedral solids, laminar structures and non-rigidly connected structures, without sacrificing a detailed understanding of the objects or their relationships to the scene.

Three dimensional object recognition is obviously a complex problem, and this research has attempted to address many issues. A skeletal exploration seemed appropriate because the use of surface data for recognition was and still is relatively untried. Thus, examining the whole problem exposed many areas for future research, some of which are being investigated as IMAGINE II is being developed. Consequently, the results presented have not solved the problem of three dimensional scene understanding, but they are a few more steps on the way.