Զѧ
ACTA AUTOMATICA SINICA
1997ꡡ236ڡVol.23No.61997




ģ͵άʶ
Ӧǿ
ժҪʵһĻģ͵άʶϵͳʶҶͼа壬ڵƣʶڵ.ϵͳʵԶģҲȶʶĳһͼԻȡ߲֪ʶȻڸ߲֪ʶָ׼ȷƥͼӦ.⣬ܱʾı()ʶķϵͳ.ʵ֤ϵͳ൱Ƚ.
ؼʡάʶ𣬹ϵͼ棬ƥ䣬Ӧ.
MODEL BASED 3-D OBJECT RECOGNITION
LIN YINGQIANGWU LIDE 
(Computer Science Department, Fudan Univ., Shanghai200433)
AbstractA model based 3-D object recognition system is realized. The system can extract the relational face-graph of the object in the origional image, then by matching the relational face-graph with the one in the modle base, the system can recognize what kind of object is in the image. If the stereo images of one scene are given, the system can accurately establish the correspondence between the two images using higher level knowledge about the scene. In addition, the characteristic-matching method has been proposed. The main idea of this method is to recognize the object by using the faces which can best represent the characteristics of the object. The robustness of the system has been proven by experiments.
Key words3-D object recognition, relational face-graph, characteristic face, characterisitc face matching, 3-D correspondence.
1
άʶǼӾҪɲ֣ʶͼʲô͵壬ͼӳĳеλúͷ.ʶĹҪй֪ʶ(model)Ϊģ͵άʶ.Ļ˼ѡȡĳЩΪԪȷԪ֮໥ϵ󣬽ʾһϵͼ.ģ;ʶϵͳʶǰõıʾϵͼ.ϵͳͨͼгȡĹϵͼģͽƥʶͼа.
Ŀǰкܶάʶоһܼף14ݣΪĽܼף23.ʵֵάʶϵͳԻҶͼΪ롢2άΪԪϵͼ.ĴڣԴʵͼеõĹϵͼ.Ϊˣƥķͨʶܱʾ()ʶ.ʹòĹϵͼֻҪܵõһ棬ϵͳпȷʶ.⣬ṩһͼԣϵͳͨʶȡĸ߲֪ʶ׼ȷƥͼӦ.ƪⲿݴ.
2ģ͵Ľֵ
2.1ģ͵Ľ
ʵֵʶϵͳĽģף2ͬ.ֻǱϵͳڽģʱÿһģ͵õ26Ųͼ.
2.2ֵ
ϵͳԳֵƣģͿֵֻĹϵͼϵͳֻʶģͿе.ģͿдŵӾӦõҪı䣬Ľܵķͨõ.
ܶÿһɵ26ŲͼĶԳԣĳЩͼõĹϵͼͼͬģģʱϵͳģͿֻĲͬĹϵͼ(ף23).⣬Щͼдԣûкܺõطӳ.ˣʶʱϵͳִĳͼ޷ȷͣӦҪûṩһͼ󣬻ģͿѡȡһ.
ĲͼУֵܽıͣʾ478ֱʾΡƽıΡΣԱڵ5ı棬ͱŹ涨ΪĿ100+(ߵĿ-5)+30.
3ͼȡϵͼ
ϵͳͼΪҶͼ󣬳ߴ256256أҶȼΪ256.ͼʶϵͳȽбԵȡͶϱӵԤ(˹ɵͼ󣬲ؽԤ).ԤͼÿһɵȦΧ.ˣҪΧ׷㷨ͼȡÿһ棬ΪϵͼĻԪ.ͳĻصΧ׷㷨׷״ΪӵΧʱִף5еΧ׷㷨(ף6).ף7ڱ߹̵Χ׷㷨׽˶ֵͼеΧ׷.׷ٳÿΧߺ󣬾ͿԽͼҳÿ.Χһֹʱ滥Ϊ.ͼʾͼӹϵǹϵͼΪͼÿһڵ()δȷڶΣڵ廹δڵ.ΪȷͣΧ߽зָϵȴõΧߵĽṹ.õÿһΧ߽ṹ󣬿áTڵȷڵ沢뱻ڵڵ棬ͬʱָܻڵ״.󣬱ȷÿͣõĹϵͼ.״ȫָıڵ棬ҪģͿй֪ʶȷ.ͼ1ʾΪһϵͼ.
   
(a)(ͼеǱı)        (b)ϵͼ(еǱ)
ͼ1ϵͼ
4άʶ
4.1ͨƥϵͼʵʶ
ϵͳʶķǽͼгȡĹϵͼϵͳģͿеĹϵͼƥ.ƥɹϵͳʶͼе壻֮ϵͳʾʶ.ϵͼƥʵͼеͼͬ.йͼͬ㷨ʮֳõģ.
4.2ƥʵάʶ
ʵͼУĴڣϵͳԵõĹϵͼΪˣʶķ.࣬ĳһSͬһһ()ϵͼг֣ĳһŹϵͼûSͬ͵棬SΪĵһ(KF1)ӸвͼеõĹϵͼֻһ͵ı(ĳһͼظ֣òͼв͵)Ϊĵڶ(KF2).
ܻʶ𷽷Ҫ˵ǣƥʶʱпܳʶ.ȽĽ.ÿһڵһSа(SǸ)ɸýڵָÿһԪԼSڸеȨ.Ȩؼ£Եһ棬ڸвͼгֵĴԸģͿĲͼгֵĴλ01֮һ.ڱ漯гĳһУڸеȨؽӽ1ٵڶг֣ÿһеȨؽС͸ԽϸߵȨԵڶ棬ȨΪ1.ÿһָȨشӴС˳ʶʱҵȨش.
󣬱ɽлʶ.һԣʶڵʶֵԽʶԿɿԽ.
ʶ㷨.
룺ͼȡĲĹϵͼ.ܵ.
ţresultΪʶֵֵΪ0PARTIΪͼеĹϵͼPARTMΪģͿе壻PARTM-GRPARTMĹϵͼ.
1)ӲĹϵͼҳ棬KF1KF2.䷽뽨ʱķһ.KF2ֻһ.
2)ҳÿһָ壬ÿһָһϣмм.
3)治ֹһмϵĽҳ.
4)if治ֹһҽգthenÿһ(PARTM)do 5)-14).
5)result1=0,ҳPARTMйϵͼÿһŹϵͼPARTM-GR do 6)-12).
6)PARTM-GR治PARTI棬result1=11;ȡһPARTM-GRת6)PARTM-GRȡ꣬ת13).
7)-11)PARTIвPARTM-GRеĶӦͬ棬result1=12result1=13PARTIвPARTM-GRеĶӦ֮ͬܽһһӦresult1=14;result1=15ܴPARTIPARTM-GRȡͼͼƥ䣬result1=16.
12)result1result,result=result1;result1=0,ȡһPARTM-GRת6)PARTM-GRȡ꣬ת13).
13)result0PARTMΪPARTIƥ壬resultֵ.
14)result=0ȡһPARTMת5)еPARTMȡ꣬ת15).
15)ҵƥ䣬ת42)result=0ת16).endif
16)if治ֹһ result=0then
17)ҳָյ棬5мϵĲ5ѡȡָΪָΪգ򷵻زʶ.
18)Բеÿһ(PARTM) do 19)-27).
19)result1=0,ҳPARTMйϵͼÿһŹϵͼPARTM-GR do 20)-24).
20)PARTM-GR治PARTI棬result1=1ȡһPARTM-GRת20)PARTM-GRȡ꣬ת25).
21)-23)PARTIвPARTM-GRеĶӦͬ棬result1=2PARTIвPARTM-GRеĶӦ֮ͬܽһһӦresult1=3.
24)result 1resultresult=result1result1=0ȡһPARTM-GRת20)PARTM-GRȡ꣬ת25).
25)result0ұಢָյı=5result=result+3.
26)result0PARTMΪPARTIƥ壬resultֵ.
27)result=0ȡһPARTMת19)еPARTMȡ꣬ת28).
28)ҵƥ䣬ת42).endif
29)ifֻһthen
30)ѡȡָΪΪգ򷵻زʶ.
31)Բеÿһ(PARTM) do 32)-41)32)-41)18)-28)ķһֻһ棬ԸKF2棬KF2棬resultֵ13.endif
42)ҵƥ䣬򷵻ͼʶֵ.
5ʵ
ô˹ɵͼһЩʵͼʶϵͳ˲ԣϵͳ˹ɵͼ൱ȽԣʵͼҲһȽ.ڱԵȡӾͲ㴦ȡýϺýϵͳʵͼȽ.
Ŀǰϵͳȷָͼ1ĳʶеÿһ.ͼ2(b)SobelԵõһʵͼȡֵõıԵͼͼ2(a)˹ĸʵͼıԵ12֮ҶȽӽĸţ֮ıԵδ.ͼе3棬ͷֱΪ33(8)31(6)8().ģͿвͬʱ壬ϵͳѡȡҪΪ33棬нָͼ2(a)е壬Ϊ33ǸKF1ȨΪ1ֻһ.ˣͼĹϵͼϵͳȷʶ˸.

(a)˹Ե(b)ʵͼıԵͼ
ͼ2
ע.ͼеСȦ(γɵļٱ)ͼͷָĶϱΧ׷ʱȥ.
1)ȻѧĿ.
߼飺Ӧǿѧѧϵ̡˶ʿ.ҪģʽʶͼͼӾо.
¡ѧѧϵڡʿʦ.ҪģʽʶӾѧԼӦõȷо.Ȼѧȶİƪ.
ߵλѧѧϵϺ200433
ο
1ݡArman F, Aggarwal J K, Model-based object recognition in dense range imagesA review.ACM Comput. Surveys, 1993.3.
2ݡDickinson S J, Pentland A P, Rosenfeld A.From volumes to views: An approach to 3-D object recognition. CVGIPl:IU 1992, 55(2):130-154.
3ݡDickinson S J, Pentland A P, Rosenfeld A. 3-D shape recovery using distributed aspect matching. IEEE Trans. PAMI, special issue on Interpretation of 3-D Scenes, Part , 1992, 14(2):174-198.
4ݡ.ӾϺѧ磬1993.
5ݡPavlidis T. Algorithms for Graphics and Image Processing, Rockville: Computer Science Press, 1982.
6ݡӦǿ.Χ׷㷨ĸĽ.ģʽʶ˹ܣ1994, 7(3):215-226.
7ݡ£Ӧǿ.ڱ߹̵Χ׷ٺΧߵṹ.ѧ199619(6)457-465.
ոڡ95-07-13
