제목:  Wavelets: Using Mathematics to Extract Detailed Information

​​​​​​​​​​​​​​​​​​연사: 허영미 교수 (연세대학교 수학과​)​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​

장소: 과학관 B102호​

Abstract:

 In this presentation, the main focus is on the introduction of wavelets, with special attention to wavelet filters and wavelet coefficients. Wavelet filters are essential for generating wavelet coefficients, yet constructing these filters can be challenging, especially when aiming to achieve specific properties. Wavelet coefficients are useful for extracting fine details from the broader context within datasets, effectively isolating highfrequency components from low-frequency ones. We will also discuss how the Laplacian pyramid algorithm, an alternative method, enables the acquisition of wavelet coefficients without the direct need for wavelet filters. If time permits, we will explore a novel method for incorporating wavelet features into deep learning algorithms.