m 
with those of the model top. In Figure 5(b), Region 41 does not 
match with the model region if one layer Hopfield neural 
network is used. However, its edges match with those of the 
model top in the line pattern layer of the two-layer Hopfield 
neural network. Region 41 is thus recognized properly. 
5. CONCLUSIONS 
In this paper, we introduced a two-layer Hopfield neural 
network for recognition of 3-D objects from AIMS imagery. In 
this approach, a 3-D object is back-projected onto an image 
with known interior and exterior orientation parameters. The 
extracted edges and regions are then matched with the back- 
projected features in a two-layer Hopfield neural network. The 
network considers the relationship of image regions and model 
regions and the similarity between image edges and model 
edges simultaneously. The experimental results of truck 
recognition showed very encouraging potential. The full result 
including truck velocity estimation is in Li et al. (1998b) at 
http://shoreline.eng.ohio-state.edu. 
6. ACKNOWLEDGMENTS 
We appreciate the Research Grants from the National Science 
Foundation (NSF CMS-9812783) and OSU Center for 
Mapping. AIMS data are provided by OSU Center for 
Mapping. 
REFERENCES 
Bulthoff, H.H., S.Y. Edelman, and M.J. Tarr, 1994, How are 
Three-dimensional Objects Represented in the Brain?, MIT 
A.I. Memo No. 1479. 
Clemens, D.T., 1991, Region-Based Feature Interpretation for 
Recognizing 3-D Models in 2-D Images, MIT AI TR-1307. 
Cohen, F.S. and J.-Y. Wang, 1994, Part I: Modeling Images 
Curves Using Invariant 3-D Object Curve Models-A Path to 3- 
D Recognition and Shape Estimation from Image Contours, 
IEEE Trans. Pattern Analysis and Machine Intelligence, 
Vol.16, No.l, pp.1-12. 
Hopfield, J.J. and D.W. Tank, 1985, "Neural" Computation of 
Decisions in Optimization Problems, Biol. Cybern., Vol. 52, 
pp.141-152. 
Horaud, R., 1987, New Methods for Matching 3-D Objects 
with Single Perspective Views, IEEE Trans. Pattern Analysis 
and Machine Intelligence, Vol.9, No.3, pp.401-412. 
Jacobs, D.W., 1992, Recognizing 3-D Objects Using 2-D 
Images, PhD Thesis, MIT Artificial Intelligence Laboratory. 
Lamdan, Y., J.T. Schwartz, and H.J. Wolfson, 1990, Affine 
Invariant Model-Based Object Recognition, IEEE Trans. 
Pattern Analysis and Machine Intelligence, Vol.6, No.5, 
pp.578-589. 
Li, R., W. Wang, and H.-Z. Tseng, 1998a, Object Recognition 
and Measurement from Mobile Mapping Image Sequences 
Using Hopfield Networks: Part I, Proceedings of ASPRS/RT 
Annual Conference 1998, CDROM. 
Li, R., F. Ma and Z. Tu, 1998b, Object Recognition from 
AIMS Data Using Neural Networks, Dept, of Civil and 
Environmental Engineering and Geodetic Science, The Ohio 
State University, http://shoreline.eng.ohio-state.edu. 
Lin, W.-C., F.-Y. Liao, C.-K. Tsao and T. Lingutla, 1991, A 
Hierarchical Multiple-View Approach to Three-Dimensional 
Object Recognition, IEEE Trans, on Neural Networks, Vol.2, 
No.l, pp.84-92. 
Nasrabadi, N.M. and Y.C. Chang, 1992, Hopfield Network for 
Stereo Vision Correspondence, IEEE Trans, on Neural 
Networks, Vol.3, No.l, pp.5-12. 
Pontil, M. and A. Verri, 1998, Support Vector Machines for 3- 
D Object Recognition, IEEE Trans. Pattern Analysis and 
Machine Intelligence, Vol.20, No.6, pp.637-646. 
Shufelt, J., 1996, Projective Geometry and Photometry for 
Object Detection and Delineation, PhD thesis, CMU-CS96- 
164. 
Suganthan, P.N., E.K. Teoh, and D.P.Mital, 1995a, Pattern 
Recognition by Graph Matching Using the Potts Mft Neural 
Networks, Pattern Recognition, Vol.28, No.7, pp.997-1009. 
Suganthan, P.N., E.K. Teoh, and D.P.Mital, 1995b, Pattern 
Recognition by Homomorphic Graph Matching Using Hopfield 
Neural Networks, Image and Vision Computing, Vol. 13, No.l, 
pp.45-60. 
Young, S.S., P.D. Scott, and M.N. Nasser, 1997, Object 
Recognition Using Multilayer Hopfield Neural Network, IEEE 
Trans, on Image Processing, Vol.6, No.3, pp.357-371. 
Ullman, S. and R. Basri, 1991, Recognition by Linear 
Combinations of Models, IEEE Trans. Pattern Analysis and 
Machine Intelligence, Vol. 13, No. 10, pp.992-1005. 
Wang, J.-Y. and F.S. Cohen, 1994, Part II: 3-D Object 
Recognition and Shape Estimation from Image Contours Using 
B-Splines, Shape Invariant Matching, and Neural Network, , 
IEEE Trans. Pattern Analysis and Machine Intelligence, 
Vol.16, No.l, pp. 13-23. 
7A-3-6