TY - GEN
T1 - Image blurring effects due to depth discontinuitites
T2 - 2nd European Conference on Computer Vision, ECCV 1992
AU - Nguyen, Thang C.
AU - Huang, Thomas S.
N1 - Funding Information:
This work is supported by the National Science Foundation under the Creativity in Engineering Award EID-8811553, and grant IRI-89-02728
PY - 1992
Y1 - 1992
N2 - A new model (called multi-component blurring or MCB) to account for image blurring effects due to depth discontinuities is presented. We show that blurring processes operating in the vicinity of large depth discontinuities can give rise to emergent image details, quite distinguishable but nevertheless un-explained by previously available blurring models. In other words, the maximum principle for scale space [Per90] does not hold. It is argued that blurring in high-relief 3-D scenes should be more accurately modeled as a multi-component process. We present results form extensive and carefully designed experiments, with many images of real scenes taken by a CCD camera with typical parameters. These results have consistently support our new blurring model. Due care was taken to ensure that the image phenomena observed are mainly due to de-focussing and not due to mutual illuminations [For89], specularity [Hea87], objects’ “finer” structures, coherent diffraction, or incidental image noises. [Gla88] We also hypothesize on the role of blurring on human depth-from-blur perception, based on correlation with recent results from human blur perception. [Hes89].
AB - A new model (called multi-component blurring or MCB) to account for image blurring effects due to depth discontinuities is presented. We show that blurring processes operating in the vicinity of large depth discontinuities can give rise to emergent image details, quite distinguishable but nevertheless un-explained by previously available blurring models. In other words, the maximum principle for scale space [Per90] does not hold. It is argued that blurring in high-relief 3-D scenes should be more accurately modeled as a multi-component process. We present results form extensive and carefully designed experiments, with many images of real scenes taken by a CCD camera with typical parameters. These results have consistently support our new blurring model. Due care was taken to ensure that the image phenomena observed are mainly due to de-focussing and not due to mutual illuminations [For89], specularity [Hea87], objects’ “finer” structures, coherent diffraction, or incidental image noises. [Gla88] We also hypothesize on the role of blurring on human depth-from-blur perception, based on correlation with recent results from human blur perception. [Hes89].
KW - Active vision
KW - Depth-from-blur
KW - Human blur perception
KW - Incoherent imaging of 3-D scenes
KW - Multi-component image blurring (MCB)
KW - Point-spread functions (kernels)
UR - http://www.scopus.com/inward/record.url?scp=85009549032&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85009549032&partnerID=8YFLogxK
U2 - 10.1007/3-540-55426-2_39
DO - 10.1007/3-540-55426-2_39
M3 - Conference contribution
AN - SCOPUS:85009549032
SN - 9783540554264
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 347
EP - 362
BT - Computer Vision - ECCV 1992 - 2nd European Conference on Computer Vision, Proceedings
A2 - Sandini, Giulio
PB - Springer
Y2 - 19 May 1992 through 22 May 1992
ER -