| Presentation | 2016-03-23 Classification of binocular disparities of a three-layered neural network with inputs of hybrid-type disparity detectors Ken Shibata, Hiroki Tanaka, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Binocular complex neurons in the primary visual cortex (V1) are successfully modeled as “hybrid-type” disparity-energy units whose receptive fields (RFs) contain both phase and position differences between the eyes. To understand the mechanisms for decoding stimulus binocular disparities using signals of these units, we constructed a three-layered neural network that included the hybrid-type units for the input layer and learned to classify binocular disparities (-0.5 to 0.5 degrees in 0.1 degree steps) in the random-dot stereograms. After learning, the hybrid-type network correctly (> 95%) classified the stimulus binocular disparities. Although the range of RF positional differences of the hybrid-type units (-0.7°~ 0.7°) is larger than that of the stimulus binocular disparities, nearly all of the middle-layer units are most strongly connected with the input-layer units whose RF phase difference is the maximum used in this experiment. The present results indicate that, for correct classification, the range of binocular disparities coded by the input layer should be considerably larger than that of stimulus binocular disparity. This may be one factor for the middle layer units being strongly connected with the hybrid-type units with large RF phase differences. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | binocular disparity / V1 / neural network / disparity energy model |
| Paper # | NC2015-89 |
| Date of Issue | 2016-03-15 (NC) |
| Conference Information | |
| Committee | MBE / NC |
|---|---|
| Conference Date | 2016/3/22(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Tamagawa University |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Tetsuo Kobayashi(Kyoto Univ.) / Toshimichi Saito(Hosei Univ.) |
| Vice Chair | Yutaka Fukuoka(Kogakuin Univ.) / Shigeo Sato(Tohoku Univ.) |
| Secretary | Yutaka Fukuoka(akita noken) / Shigeo Sato(Kogakuin Univ.) |
| Assistant | Takenori Oida(Kyoto Univ.) / Ryota Horie(Shibaura Inst. of Tech.) / Hiroyuki Kanbara(Tokyo Inst. of Tech.) / Hisanao Akima(Tohoku Univ.) |
| Paper Information | |
| Registration To | Technical Committee on ME and Bio Cybernetics / Technical Committee on Neurocomputing |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Classification of binocular disparities of a three-layered neural network with inputs of hybrid-type disparity detectors |
| Sub Title (in English) | |
| Keyword(1) | binocular disparity |
| Keyword(2) | V1 |
| Keyword(3) | neural network |
| Keyword(4) | disparity energy model |
| 1st Author's Name | Ken Shibata |
| 1st Author's Affiliation | Kyoto Sangyo University(Kyoto Sangyo Univ.) |
| 2nd Author's Name | Hiroki Tanaka |
| 2nd Author's Affiliation | Kyoto Sangyo University(Kyoto Sangyo Univ.) |
| Date | 2016-03-23 |
| Paper # | NC2015-89 |
| Volume (vol) | vol.115 |
| Number (no) | NC-514 |
| Page | pp.pp.113-117(NC), |
| #Pages | 5 |
| Date of Issue | 2016-03-15 (NC) |