{"id":40310,"date":"2020-04-09T20:45:54","date_gmt":"2020-04-09T20:45:54","guid":{"rendered":"http:\/\/www.labri.fr\/perso\/barla\/blog\/?p=40310"},"modified":"2020-04-09T20:45:56","modified_gmt":"2020-04-09T20:45:56","slug":"vision-theories-w-alex","status":"publish","type":"post","link":"https:\/\/www.labri.fr\/perso\/barla\/blog\/?p=40310","title":{"rendered":"Vision theories w\/ Alex"},"content":{"rendered":"<p id=\"top\" \/>\n<p>One aspect that is always bugging me in psycho-physics is that the<br>experimenter\/scientist decides a priori what are the potentially<br>relevant visual properties.<br>Well, one has to start somewhere, and it makes sense to talk about shape<br>and material ; but I always wondered whether a different set of<br>properties could be more relevant?<br>You can draw a parallel to changes of variable in mathematics: an<br>intractable problem can become much simpler to solve once you consider a<br>different set of variables!<\/p>\n\n\n\n<p class=\"has-text-align-right\">Yes, I <strong><em>totally<\/em> <\/strong>agree with this. For the proposal I had to put something that sounded concrete, but I would much rather do something more exploratory, and figure out what exactly the brain is processing \/ &#8220;cares about&#8221;. I don&#8217;t believe it is neatly organised into processing &#8220;shape&#8221;, &#8220;surface material&#8221;, &#8220;motion&#8221;, &#8220;colour&#8221; etc. I could potentially use my proposed stimulus sets to show that actually it doesn&#8217;t make much sense to try to isolate the processing of these different properties that experimenters come up with. I had a lot of these thoughts while writing the proposal.<\/p>\n\n\n\n<p>You ask &#8220;what&#8221;, &#8220;where&#8221;, &#8220;when&#8221; and &#8220;how&#8221; in the<br>document. But what about &#8220;why&#8221; or its cousin (only in English) &#8220;how come&#8221;?<br>For instance, &#8220;why is it that these image features characterize<br>different material classes&#8221;? (objective 1) ; a potential answer could<br>reside in ecological optics again.<br>Or, even more difficult, &#8220;how come the brain evolved to process visual<br>signals that way&#8221; (objectives 2-4); here one starts to touch on (strong)<br>AI questions.<\/p>\n\n\n\n<p class=\"has-text-align-right\">Interesting questions. Roland Fleming would say that images vary in characteristic ways and we learn these statistical regularities and then can use them to discriminate between categories (we learn the image features because they are there, and then can use them for stuff). If I understand what you mean by ecological optics, then this is more the approach taken by Bart Anderson and Phil Marlow &#8211; they look at the physical differences between classes of materials (say how matte and glossy things reflect light), and then see how that translates to differences in images. I think that is similar to your potential and effective cues. But Bart and Phil seem to be in the group of psychophysicists who a priori decide what the relevant perceptual dimensions are (glossy vs. matte, translucent vs. opaque). They get away with it because of the specific types of glossy\/translucent stimuli they look at. I think I lie somewhere in between Roland and Bart&#8217;s views. I don&#8217;t think our perceptual experience can be neatly divided into these dimensions, but I also don&#8217;t fully understand Roland&#8217;s views and how that fits with resolving 3D shape (and how it seems to be simultaneously resolved along with material).<\/p>\n\n\n\n<p>Until recently, I also thought that the &#8220;truth&#8221; should be inbetween these two views; but maybe now I&#8217;d rather say that a good theory of vision should encompass both approaches. On one side, we must consider how the &#8220;physical world&#8221; manifests itself in images, and I like to think about it in successive steps: from non-observable physical properties (say microscopic structures) to potential cues (e.g., BRDF components) to effective cues (like highlights distorted by shape) in the image. On the other side, you have a visual system that has to make use of relevant information in light coming to his eyes, as quickly and effortlessly as possible; for this it will rely on what&#8217;s available of course, which is to say effective cues, but it will also have to take different shortcuts that may vary depending on many factors (current mood, urge, visual target, specific task, experience, etc). In this latter case, a statistical approach becomes useful; but it HAS to be (even loosely) connected to effective cues and through it potential cues, &#8220;down&#8221; to the physical world, to make ecological relevance ; otherwise the organism would probably not survive according to natural selection!<br><br>I&#8217;ve actually went a step further in &#8220;subjectivity&#8221; recently, after reading Jakob von Uexkull&#8217;s book &#8220;A Stroll Through the Worlds of Animals and Men: A Picture Book of Invisible Worlds&#8221;, where he introduces the idea of Umwelt. To say it simply, each different animal has its own world or Umwelt, and there is no objective world that needs to be considered. That&#8217;s why I&#8217;ve put &#8220;physical world&#8221; in quotes above: I consider this physical account is also in part subjective. For instance, what we call a material depends on the scale at which it is observed, the wavelengths we consider depend on the properties of human eyes, etc. So there&#8217; s not physics on one side and and vision on the other side; they are tied together right from the beginning. This is a more metaphysical question, so maybe it&#8217; s not that important in the end<br><br>In any case, what I think is missing is a way to TEST different theories, ways to design experiments that tease apart which one is right (if there&#8217;s any). I don&#8217;t think any existing theory about the perception of complex visual stimuli is mature enough to be fully testable.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>One aspect that is always bugging me in psycho-physics is that theexperimenter\/scientist decides a priori what are the potentiallyrelevant visual properties.Well, one has to start somewhere, and it makes sense to talk about shapeand material ; but I always wondered whether a different set ofproperties could be more relevant?You can draw a parallel to changes &#8230; <a title=\"Vision theories w\/ Alex\" class=\"read-more\" href=\"https:\/\/www.labri.fr\/perso\/barla\/blog\/?p=40310\" aria-label=\"Read more about Vision theories w\/ Alex\">Read more<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[21],"tags":[],"class_list":["post-40310","post","type-post","status-publish","format-standard","hentry","category-discuss"],"_links":{"self":[{"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/posts\/40310","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=40310"}],"version-history":[{"count":1,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/posts\/40310\/revisions"}],"predecessor-version":[{"id":40311,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=\/wp\/v2\/posts\/40310\/revisions\/40311"}],"wp:attachment":[{"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=40310"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=40310"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.labri.fr\/perso\/barla\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=40310"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}