Detection of unknown object from dynamic scene with known moving object

There are many methods to extract moving objects in a scene using background subtraction. However, most methods assume that there are no moving objects except intruders in the observing space. Therefore, we propose the iterative optimal projection method to estimate a varied background in real time from a dynamic scene with intruders. At first, the background images are collected for a while, because we assume that the motion of background is well known. Then, the background images are compressed using eigenspace method to form a database. While the monitoring of the scene, new image is taken by a camera, and the image is projected onto the eigenspace to estimate the background. Nevertheless, the estimated image is much affected by the intruders, so that the region of intruder is calculated by using background subtraction using former estimated background. Then eigenspace projection is executed using the image whose region of the intruder is replaced by the former background and we have updated background. We proved that the cycle converges to a correct background image and we confirmed we can calculate the right region of the object through some experiments.

Distinguish human from door motion video 1 : Input image with known moving object(door of lift) and unknown object(human) video 2 : Generated background image sequence video 3 : Extracted unknown object region

Performed with :

See also :

• Satoshi Kawabata, Shinsaku Hiura, Kosuke Sato : Real-Time Extraction of Unknown Region from Dynamic Background by Robust Projection onto Eigen Space, IEICE Trans. on Information and systemsCVol. J89-D, No. 4, pp.826-835 (in Japanese) (Apr. 2006)
  

• Satoshi Kawabata, Shinsaku Hiura, Kosuke Sato : Real-time Background Estimation from Occluded Image Using Iterative Optimal Projection Method, SSR2005 CD-ROM (Mar. 2005 at San Francisco)
  
• Satoshi Kawabata, Shinsaku Hiura, Kosuke Sato : Unknown Object Detection based on Trained Background Database, SICE2005 (Aug. 2005)
  
• Satoshi Kawabata, Shinsaku Hiura, Kosuke Sato : Unknown Object Detection based on Trained Background Database, SSR2006 CD-ROM(Mar. 2006 at Vienna)
  
• Satoshi Kawabata, Shinsaku Hiura and Kosuke Sato : Real-Time Detection of Anomalous Objects in Dynamic Scene, Proc. of 18th ICPR, pp.1171-1174 (Aug. 2006 at Hong Kong)