Directional multiscale representations and applications in digital neuron reconstruction
Directional multiscale representations and applications in digital neuron reconstruction
复制标题
DOI：
10.1016/j.cam.2018.09.003
复制
发表时间：
2019-03-15
影响因子：
2.4
通讯作者：
Labate，Demetrio
中科院分区：
文献类型：
--
作者：
Kayasandik，Cihan;Guo，Kanghui;Labate，Demetrio
在多尺度表示领域的最新进展已经刺激了新一代的强大技术的出现，用于有效地分析图像和其他多维数据。这些新技术使量化的基本几何特征，在复杂的成像数据，从而改进算法的图像恢复，特征提取和分类。我们讨论了这些想法在神经科学成像中的应用，并描述了一种新的方法，用于准确和有效地识别多细胞图像中的神经元细胞体。该方法有助于设计一种新的神经元跟踪算法。
Recent advances in the field of multiscale representations have spurred the emergence of a new generation of powerful techniques for the efficient analysis of images and other multidimensional data. These novel techniques enable the quantification of essential geometric characteristics in complex imaging data resulting in improved algorithms for image restoration, feature extraction and classification. We discuss the application of these ideas in neuroscience imaging and describe a novel method for the accurate and efficient identification of cellular bodies of neurons in multicellular images. This method is instrumental to the design of a novel algorithm for neuronal tracing.