Hi, I have heard you tried using CLA in vision but it dint go so well. Wouldn't it be productive if an image is passed through retina, lgn and then into CLA instead of directly using raw image on cla. I have been working on modelling retina and simulating it to check its feasibility. If it works it would output patterns like our optic nerve. There would be motion selective cells, color information, boundary information from center surround cells. I am still stuck on horizontal cells, because in invertebrates they have color opponency which makes it easy for contrast enhancement in bipolars but in primates there is no color opponency. We just have hyperpolarization in our horizontal cells as light falls on photo receptors. Since primates have better vision than others I know this hyperpolarization must be for a reason but i dont know what. So kind of stuck.
Jeff hawkins responses on neuroscience are so deep and interesting so I thought of bringing this up. Would love your opinions as always ! Aseem Hegshetye _______________________________________________ nupic mailing list [email protected] http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
