SDNet: A Versatile Squeeze-and-Decomposition Network for Real-Time Image Fusion

Zhang, Hao; Ma, Jiayi

doi:10.1007/s11263-021-01501-8

Cited by 256 publications

(67 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Qualitative comparison of our CEFusion with seven state‐of‐the‐art methods on two typical CT and MRI image pairs. From left to right: CT and MRI‐T1/MRI‐T2 images, fusion results of TDS [1], PAPCNN [2], LRD [4], DSAGAN [6], PMGI [8], SDNet [9], MSDNet [11] and our CEFusion. For more intuitive comparison, some regions are enlarged and shown in the bottom corners…”

Section: Resultsmentioning

confidence: 99%

“…Qualitative comparison of our CEFusion with seven state‐of‐the‐art methods on other types image pairs. From left to right: MRI and PET/SPECTtc/SPECTti images, fusion results of TDS [1], PAPCNN [2], LRD [4], DSAGAN [6], PMGI [8], SDNet [9], MSDNet [11] and our CEFusion…”

Section: Resultsmentioning

confidence: 99%

“…In this section, we compare CEFusion with four of state‐of‐the‐art methods deep learning‐based methods and three traditional methods, Deep learning‐based methods include MSDNet [11], PMGI [8], SDNet [9], and DSAGAN [6]. In addition, the three traditional methods applied for comparison include PAPCNN [2], TDS [1], LRD [4].…”

Section: Methodsmentioning

confidence: 99%

“…Zhang et al. designed an adaptive decision block to constrain gradient information to achieve real‐time multi‐mode digital photography image fusion (SDNet) [9]. In [43], Zhang et al.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

CEFusion: Multi‐Modal medical image fusion via cross encoder

Zhu

Wang

Chen

et al. 2022

IET Image Processing

View full text Add to dashboard Cite

Most existing deep learning-based multi-modal medical image fusion (MMIF) methods utilize single-branch feature extraction strategies to achieve good fusion performance. However, for MMIF tasks, it is thought that this structure cuts off the internal connections between source images, resulting in information redundancy and degradation of fusion performance. To this end, this paper proposes a novel unsupervised network, termed CEFusion. Different from existing architecture, a cross-encoder is designed by exploiting the complementary properties between the original image to refine source features through feature interaction and reuse. Furthermore, to force the network to learn complementary information between source images and generate the fused image with high contrast and rich textures, a hybrid loss is proposed consisting of weighted fidelity and gradient losses. Specifically, the weighted fidelity loss can not only force the fusion results to approximate the source images but also effectively preserve the luminance information of the source image through weight estimation, while the gradient loss preserves the texture information of the source image. Experimental results demonstrate the superiority of the method over the state-of-the-art in terms of subjective visual effect and quantitative metrics in various datasets.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

“…Zhang et al. designed an adaptive decision block to constrain gradient information to achieve real‐time multi‐mode digital photography image fusion (SDNet) [9]. In [43], Zhang et al.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

CEFusion: Multi‐Modal medical image fusion via cross encoder

Zhu

Wang

Chen

et al. 2022

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…The training set is composed of 27 264 pairs of image patches with size . Two test image pairs are chosen from RGB-NIR dataset for comparing with different methods qualitatively and another 20 test image pairs for quantitative We conduct the comparison experiments with seven state-of-the-art fusion methods including VSM [10], CVN [11], WLP [9], CNI [12], GF [13], U2fusion [6] and SDNet [14]. The parameters are set as follows: training epoch is set to 20, and number of batch images is set to 12.…”

Section: × 64mentioning

confidence: 99%