https://scholars.lib.ntu.edu.tw/handle/123456789/607209
標題: | DesmokeNet: A Two-stage Smoke Removal Pipeline Based on Self-Attentive Feature Consensus and Multi-Level Contrastive Regularization | 作者: | Chen W Lou H Fang H Chen I Chen Y JIAN-JIUN DING SY-YEN KUO |
關鍵字: | convolutional neural network;multi-level contrastive regularization;pixel compensation;self-attentive feature consensus;Two-stage smoke removal;Deep neural networks;Large dataset;Pipelines;Pixels;Color distortions;Compensation network;Discriminative ability;High-level visions;Large-scale dataset;Non-homogeneous distribution;Single images;State-of-the-art methods;Smoke | 公開日期: | 2021 | 來源出版物: | IEEE Transactions on Circuits and Systems for Video Technology | 摘要: | In image processing, smoke may degrade visibility and deteriorate the performance of high-level vision applications. Therefore, single image smoke removal is crucial for computer vision. Currently, existing smoke removal algorithms mainly leverage handcrafted priors. Moreover, these methods usually apply haze removal methods to perform smoke removal due to the similarity between smoke and haze. However, these methods cannot sufficiently address the degradation of thick smoke and may suffer from residual smoke and color distortion problems due to the non-global and non-homogeneous distribution of smoke. In this paper, to solve the aforementioned problems, an end-to-end deep neural network called DesmokeNet is proposed. We construct a two-stage recovered pipeline to remove the smoke in different thicknesses. The light and thick smoke is first removed locally by the smoke removal network (SRN). The missing pixels in the thick smoke are then recovered by the pixel compensation network (PCN). Moreover, we proposed the thickness-aware pixel loss and the dark channel loss to suppress the residual smoke. To further increase the discriminative ability of the DesmokeNet, we proposed self-attentive feature consensus loss and multi-level contrastive regularization loss to improve the performance of smoke removal. Finally, to train the proposed method, we construct the first large-scale dataset containing synthetic and real-world data. Extensive experiments show that the proposed method outperforms favorably against other state-of-the-art methods quantitatively and qualitatively. IEEE |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85113267421&doi=10.1109%2fTCSVT.2021.3106198&partnerID=40&md5=f997e48d4fac09b8c271b3cb7367b855 https://scholars.lib.ntu.edu.tw/handle/123456789/607209 |
ISSN: | 10518215 | DOI: | 10.1109/TCSVT.2021.3106198 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。