Feature Reduction and Noise Removal in SURF Framework for Efficient Object Recognition in Images
Speeded up Robust Features (SURF) is an interest point detector and descriptor which has been popularly used for object recognition. However, in real time object recognition applications, SURF framework can not be used because of its expensive nature. In this paper, a feature reduction process is proposed by using only the most repeatable features for matching. The feature reduction step results in a remarkable computational speed up with little loss of accuracy. A noise-reduction process allows a further increase in matching speed and also reduces the false positive rates. A modified definition of the second-neighbor in the nearest neighbor ratio matching strategy allows matching with increased reliability. The comparative analysis with SURF framework shows that the proposed framework can be useful in applications where the accuracy can be sacrificed to save computational cost.
KeywordsObject recognition SURF features Pose estimation Noise reduction Feature matching
This research was supported by: (1)The Industrial Strategic technology development program, 10041772, (The Development of an Adaptive Mixed-Reality Space based on Interactive Architecture) funded by the Ministry of Knowledge Economy(MKE, Korea), (2) The MKE(The Ministry of Knowledge Economy), Korea, under IT/SW Creative research program supervised by the “NIPA(National IT Industry Promotion Agency)” (NIPA-2012- H0502-12-1013).
- 1.Ejaz N, Tariq TB, Baik SW (2012) Adaptive key frame extraction for video summarization using an aggregation mechanism. J Vis Commun Image Represent 23(7):1031–1040Google Scholar
- 2.Ejaz N, Manzoor U, Nefti S, Baik SW (2012) A collaborative multi-agent framework for abnormal activity detection in crowded areas. Int J Innov Comput Inform Control 8(6):4219–4234Google Scholar
- 3.Ejaz N, Lee JW, Kim W, Lim C, Joo S, Baik SW (2012) Automated selection of appropriate advertisements for digital signage by analyzing crowd demographics, Special issue on computer convergence technologies. Inform Int Interdiscip J 15(5):2019–2030Google Scholar
- 4.Bay H, Tuytelaars T, Gool LV (2008) Speeded-up robust features (SURF). Comp Vis Image Underst 110(3):346–359Google Scholar
- 5.Lowe DG (1999) Object recognition from local scale-invariant features, Proceedings of the international conference on computer vision, pp 1150–1157Google Scholar
- 6.Lowe DG (2004) Distinctive image features from scale-invariant key points. Int J Comp Vis 60(2):91–110Google Scholar
- 7.Lee S, Kim K, Kim JY, Kim M, Yoo HJ (2010) Familiarity based unified visual attention model for fast and robust object recognition. Pattern Recognit 43(3):1116–1128Google Scholar
- 8.MacQueen JB (1967) Some methods for classification and analysis of multivariate observations. In: Cam LML, Neyman J (eds) In proceeding of the Berkeley symposium on mathematical statistics and probability, vol 1. University of California Press, California, pp 281–297Google Scholar
- 9.Nistér D, Stewénius H (2006) Scalable recognition with a vocabulary tree. In: Proceeding of IEEE conference on computer vision and pattern recognition, vol 2. pp 2161–2168Google Scholar
- 10.Bashir F, Porikli F (2006) Performance evaluation of object detection and tracking systems. In: Proceeding of international workshop on performance evaluation of tracking and surveillanceGoogle Scholar
- 11.Lin L, Wu T, Porway J, Xu Z (2009) A stochastic graph grammar for compositional object representation and recognition. Pattern Recognit 42(7):1297–1307Google Scholar