LEARN! Seminar Noor Seijdel
12.30 - 13.30
Join Zoom Meeting https://vu-live.zoom.us/j/93805427201 Passcode: 488688 This presentation will be recorded (not Q&A)
Investigating visual processing using computational models
Congres / symposium / seminar
One way to understand how the human brain processes visual information involves building computational models that account for human-level behavior under different conditions. Such models become even more interesting when the underlying anatomy, architecture or ‘physiology’ can be linked to the behavior of interest. In this talk, I will first present some of the work that I did during my PhD, in which we compared performance of different deep convolutional neural networks to human behavior to investigate scene segmentation during visual object recognition. Then I will discuss more recent work in which we investigate OB1-reader, a reading-model that simulates the processes behind reading in the human brain, by simulating human behavioral performance and evoked EEG activity during word recognition tasks.