Lei Xiang, Yang Li, Weili Lin, Qian Wang, Dinggang Shen
{"title":"非配对深度交叉模态合成与快速训练。","authors":"Lei Xiang, Yang Li, Weili Lin, Qian Wang, Dinggang Shen","doi":"10.1007/978-3-030-00889-5_18","DOIUrl":null,"url":null,"abstract":"<p><p>Cross-modality synthesis can convert the input image of one modality to the output of another modality. It is thus very valuable for both scientific research and clinical applications. Most existing cross-modality synthesis methods require large dataset of paired data for training, while it is often non-trivial to acquire perfectly aligned images of different modalities for the same subject. Even tiny misalignment (i.e., due patient/organ motion) between the cross-modality paired images may place adverse impact to training and corrupt the synthesized images. In this paper, we present a novel method for cross-modality image synthesis by training with the unpaired data. Specifically, we adopt the generative adversarial networks and conduct the fast training in cyclic way. A new structural dissimilarity loss, which captures the detailed anatomies, is introduced to enhance the quality of the synthesized images. We validate our proposed algorithm on three popular image synthesis tasks, including brain MR-to-CT, prostate MR-to-CT, and brain 3T-to-7T. The experimental results demonstrate that our proposed method can achieve good synthesis performance by using the unpaired data only.</p>","PeriodicalId":92501,"journal":{"name":"Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support : 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, held in conjunction with MICCAI 2018, Granada, Spain, S...","volume":"11045 ","pages":"155-164"},"PeriodicalIF":0.0000,"publicationDate":"2018-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/978-3-030-00889-5_18","citationCount":"18","resultStr":"{\"title\":\"Unpaired Deep Cross-Modality Synthesis with Fast Training.\",\"authors\":\"Lei Xiang, Yang Li, Weili Lin, Qian Wang, Dinggang Shen\",\"doi\":\"10.1007/978-3-030-00889-5_18\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>Cross-modality synthesis can convert the input image of one modality to the output of another modality. It is thus very valuable for both scientific research and clinical applications. Most existing cross-modality synthesis methods require large dataset of paired data for training, while it is often non-trivial to acquire perfectly aligned images of different modalities for the same subject. Even tiny misalignment (i.e., due patient/organ motion) between the cross-modality paired images may place adverse impact to training and corrupt the synthesized images. In this paper, we present a novel method for cross-modality image synthesis by training with the unpaired data. Specifically, we adopt the generative adversarial networks and conduct the fast training in cyclic way. A new structural dissimilarity loss, which captures the detailed anatomies, is introduced to enhance the quality of the synthesized images. We validate our proposed algorithm on three popular image synthesis tasks, including brain MR-to-CT, prostate MR-to-CT, and brain 3T-to-7T. The experimental results demonstrate that our proposed method can achieve good synthesis performance by using the unpaired data only.</p>\",\"PeriodicalId\":92501,\"journal\":{\"name\":\"Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support : 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, held in conjunction with MICCAI 2018, Granada, Spain, S...\",\"volume\":\"11045 \",\"pages\":\"155-164\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2018-09-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://sci-hub-pdf.com/10.1007/978-3-030-00889-5_18\",\"citationCount\":\"18\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support : 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, held in conjunction with MICCAI 2018, Granada, Spain, S...\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1007/978-3-030-00889-5_18\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2018/9/20 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support : 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, held in conjunction with MICCAI 2018, Granada, Spain, S...","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/978-3-030-00889-5_18","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2018/9/20 0:00:00","PubModel":"Epub","JCR":"","JCRName":"","Score":null,"Total":0}
Unpaired Deep Cross-Modality Synthesis with Fast Training.
Cross-modality synthesis can convert the input image of one modality to the output of another modality. It is thus very valuable for both scientific research and clinical applications. Most existing cross-modality synthesis methods require large dataset of paired data for training, while it is often non-trivial to acquire perfectly aligned images of different modalities for the same subject. Even tiny misalignment (i.e., due patient/organ motion) between the cross-modality paired images may place adverse impact to training and corrupt the synthesized images. In this paper, we present a novel method for cross-modality image synthesis by training with the unpaired data. Specifically, we adopt the generative adversarial networks and conduct the fast training in cyclic way. A new structural dissimilarity loss, which captures the detailed anatomies, is introduced to enhance the quality of the synthesized images. We validate our proposed algorithm on three popular image synthesis tasks, including brain MR-to-CT, prostate MR-to-CT, and brain 3T-to-7T. The experimental results demonstrate that our proposed method can achieve good synthesis performance by using the unpaired data only.