BROWSE
Detail
Journal Article
서브-모달리티 주의 기반 효율적인 의미론적 분할 기술
- Issue Date
- 2025-11
- Citation
- 한국자동차공학회 논문집, v.33, no.11, pp.957-965
- ISSN
- 1225-6382
- Publisher
- 한국자동차공학회
- Language
- Korean
- Type
- Journal Article
- Abstract
- Semantic segmentation plays a crucial role in autonomous driving by assigning pixel-wise labels to images. Traditional convolutional neural networks (CNNs) based semantic segmentation approaches incorporate conventional down-sampling layers in the initial stage to enhance computational efficiency. The feature maps in the initial layers of a CNN are more effective when individual channels capture diverse and complementary information. Thus, this work introduced a sub-modality attention network that explicitly separates high-frequency and low-frequency components, focused on integrating separated pieces of information, allowing them to complement each other's deficiencies at the early feature extraction stage. Our results demonstrate that deepening CNNs is not the only path to performance improvement-incorporating handcrafted priors, such as the Wavelet transform, can also yield significant gains.
- KSP Keywords
- Computational Efficiency, Convolution neural network(CNN), Down-sampling, Feature extractioN, Feature map, Frequency components, High frequency(HF), Initial stage, Low frequency, Semantic segmentation, Wavelet transform(DWT)
This work is distributed under the term of Creative Commons License (CCL)
(CC BY NC)
(CC BY NC)
