Model of human visual-motion sensing

...Under certain conditions, correlation-type motion detectors become formally equivalent (see Hildreth & Koch, 1987) to a variety of luminance-based motion detectors (Adelson & Bergen, 1985; Van Santen & Sperling, 1985; Watson & Ahumada, 1985)...

...[3] is not a generic VQA algorithm and targets video coding applications, where models of visual motion sensors developed in [8] are utilized to perform computations that signal the direction of motion...
...Assuming that this patch undergoes translation with a velocity where and denote velocities along the and directions respectively, the resultingvideosequenceisgivenby . Then, , the Fourier transform of , lies entirely within a plane in the frequency domain [8]...
...Translational motion has an easily accessible representation in the frequency domain and these ideas have been used to build motion estimation algorithms for video [8], [57], [58]...
...Under this assumption, the reference and test videos and consist of local image patches (such as in the example above) translating to create spatio-temporal video patches (such as ). Observe that (12) and (13) assume infinite translation of the image patches [8], which is not practical...

...In terms of the dorsal pathway that is involved in sensori-motor processing [27], computational modeling has focused on the aspect of motion perception [1, 67] and, more recently, also on attentional selection by motion [4, 6, 30, 37]...

Keywords (1)


	The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.


	BibTex \| RIS \| RefWorks Download