 |
||
 |
|
|
What's New | Greetings | Schedule | Members | Contact | Japanese |
||
|
|
|
After March 20202020MarchFebruaryJanuary2019DecemberNovemberOctoberSeptemberAugustJulyJuneMayAprilMarchFebruaryJanuary2018DecemberNovemberOctoberSeptemberAugustJulyJuneMayAprilMarchFebruaryJanuary2017DecemberNovemberOctoberSeptemberAugustJulyJuneMayAprilMarchFebruaryJanuary2016DecemberNovemberOctoberSeptemberAugustJulyJuneMayAprilBefore April 2016 |
Schedule in April 2018The 39th Perceptual Frontier Seminar: Motion, Model, and PerceptionDate and time: Monday, 16 April 2018, 16:00-18:00 Program1. Method of image feature extraction using motion information in natural movies and its application to optical flow estimation 2. The oscillating potential model of visually induced vection A new model of responding to visually induced vection was developed. First, a mathematical model was constructed based on well-documented characteristics of vection and human behavioural responses to this illusion. Then 10,000 virtual trial simulations using this Oscillating Potential Vection Model (OPVM) were conducted and the OPVM output was found to compare favorably to the empirically obtained vection data. 3. Perceptual restoration of interrupted speech and interrupted locally time-reversed speech The effects of periodical interruption with silence or noise on intelligibility of speech and locally time-reversed speech in Japanese sentences were investigated. A pilot experiment was conducted with eight to ten Japanese normal-hearing participants, using stimuli that were interrupted with 20-120 or 20-220 ms segments. The results showed that noise insertion improved intelligibility in both interrupted speech and interrupted locally-time reversed speech with intense noise; however, with less intense noise, only the intelligibility of interrupted locally-time reversed speech improved by the noise insertion. 4. Intelligibility of English mosaic speech: Influence of manipulating mosaic block duration The main purpose of this study is to observe the influence of manipulating mosaic-block duration (20-320 ms) on the intelligibility of mosaic speech. The original speech samples were transformed into mosaic by dividing the spectrogram of speech signal into blocks and after that averaging the sound energy density within each block of spectrogram period. The results of two pilot experiments showed that the intelligibility was higher than 50% when mosaic-block duration ranged from 20-80 ms, and decreased for larger duration.
We will get together after the Seminar. |
||
|
|
|
|
What's New | Greetings | Schedule | Members | Contact | Japanese |
||
|
|
|
Last updated: |
||
|
|
|