[ICJAI 2023] Acoustic NLOS Imaging with Cross-Modal Knowledge Distillation - Ui-Hyeon Shin, Seungwoo Jang and Kwangsu Kim
- 인공지능융합 연구실
- 조회수685
- 2023-07-12
Title:
Acoustic NLOS Imaging with Cross-Modal Knowledge Distillation
Author:
Ui-Hyeon Shin,
Seungwoo Jang,
Kwangsu Kim
Abstract
Acoustic non-line-of-sight (NLOS) imaging aims to reconstruct hidden scenes by analyzing reflec- tions of acoustic waves. Despite recent develop- ments in the field, existing methods still have lim- itations such as sensitivity to noise in a physical model and difficulty in reconstructing unseen ob- jects in a deep learning model. To address these limitations, we propose a novel cross-modal knowl- edge distillation (CMKD) approach for acoustic NLOS imaging. Our method transfers knowl- edge from a well-trained image network to an au- dio network, effectively combining the strengths of both modalities. As a result, it is robust to noise and superior in reconstructing unseen objects. Additionally, we evaluate real-world datasets and demonstrate that the proposed method outperforms state-of-the-art methods in acoustic NLOS imag- ing. The experimental results indicate that CMKD is an effective solution for addressing the limita- tions of current acoustic NLOS imaging methods. Our code, model, and data are available at https: //github.com/shineh96/Acoustic-NLOS-CMKD.
- 다음글
- 다음글이 없습니다.