Object detection using multi-resolution mosaic in image sequences

Object detection and video coding using background mosaic plane

“…Among them, methods based on the difference picture ͑DP͒ [1][2][3][4][5][6][7][8][9][10][11][12][13][14] have been extensively developed because they satisfy almost all requirements for object detector, viz., robustness to noise, adaptability to illumination changes, and fast detection. Background subtraction, [1][2][3][4][5][6][7][8][9] which uses DP between the current and the background image, is the most popular such method.…”

“…Background subtraction, [1][2][3][4][5][6][7][8][9] which uses DP between the current and the background image, is the most popular such method. We proposed a unified framework for background subtraction, 13 which is made up of three criteria. Two of them are about the feature and distance metrics, respectively, and the third is about the rule for adaptation of the background model to illumination changes.…”

“…So there have been many researches to solve problems in detecting objects in image sequences. [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] In general object detector should be robust to some noise and has to be adaptive to illumination changes in real environments. It is also very important to detect objects as fast as possible in realtime applications.…”

“…Also, in order to get more accurate parameters in case of outliers ͑for example, due to moving objects͒, we use outlier rejection with two stop criteria during the iteration. 13 But although the GIC is estimated correctly, the compensation error is not negligible for object detection and video coding when there is drastic GIC. We prove that in the next section.…”

Combination of Sparse Scan and Dense Scan for Fast Vision-based Object Recognition

“…This approach also requires complex and expensive camera system to operate. Other multi-resolution approaches including Ma's Ma & Staunton (2005), Walther's Walther et al (2005), and Cho's Cho & Kim (2005) are either too computational costly or not built to reduce system's computational cost. From a different point of view, part-based object detection has been investigated for applications such as human detection Mohan et al (2001); Wu & Nevatia (2007) or car detection Agarwal et al (2004).…”