Visual Analytics for model-based medical image segmentation: Opportunities and challenges

作者：

Highlights：

• We address model-based medical image segmentation from the Visual Analytics perspective.

• Expert segmentations are costly and time consuming, current automatic segmentations are not precise enough. Better segmentation algorithms are needed; they can be created with help of Visual Analytics tools.

• We identify four stages of the modeling process and present Visual Analytics methods for them.

• They improve the final result by better setting of the algorithm parameters, analysis of the input data set, and allowing visual assessment of output quality on local level.

• Based on these first results, we discussed what we see as most relevant challenges for the community.

摘要

•We address model-based medical image segmentation from the Visual Analytics perspective.•Expert segmentations are costly and time consuming, current automatic segmentations are not precise enough. Better segmentation algorithms are needed; they can be created with help of Visual Analytics tools.•We identify four stages of the modeling process and present Visual Analytics methods for them.•They improve the final result by better setting of the algorithm parameters, analysis of the input data set, and allowing visual assessment of output quality on local level.•Based on these first results, we discussed what we see as most relevant challenges for the community.

论文关键词：Medical imaging,Medical modeling,Visual Analytics,Statistical shape models

论文评审过程：Available online 13 March 2013.

论文官网地址：https://doi.org/10.1016/j.eswa.2013.03.006