Image Segmentation for Connectomics Using Machine Learning

Taşdizen, Tolga; Seyedhosseini, Mojtaba; Liu, Ting; Jones, Cory; Jurrus, Elizabeth

doi:10.1007/978-1-4614-7245-2_10

Cited by 1 publication

(2 citation statements)

References 85 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A variety of machine learning approaches have been explored for the 3D reconstruction of neurons, a problem that can be formulated as image segmentation or boundary detection [6,7]. This paper focuses on neuronal boundary detection in images from serial section EM, the most widespread kind of 3D EM [8].…”

Section: Introductionmentioning

confidence: 99%

“…Due to these issues with the z direction of the image stack [7,9], most existing analysis pipelines begin with 2D processing and only later transition to 3D. The stages are: (1) neuronal boundary detection within each 2D image, (2) segmentation of neuron cross sections within each 2D image, and (3) 3D reconstruction of individual neurons by linking across multiple 2D images [2,10].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Recursive Training of 2D-3D Convolutional Networks for Neuronal Boundary Detection

Lee¹,

Zlateski²,

Vishwanathan³

et al. 2015

Preprint

View full text Add to dashboard Cite

Efforts to automate the reconstruction of neural circuits from 3D electron microscopic (EM) brain images are critical for the field of connectomics. An important computation for reconstruction is the detection of neuronal boundaries. Images acquired by serial section EM, a leading 3D EM technique, are highly anisotropic, with inferior quality along the third dimension. For such images, the 2D maxpooling convolutional network has set the standard for performance at boundary detection. Here we achieve a substantial gain in accuracy through three innovations. Following the trend towards deeper networks for object recognition, we use a much deeper network than previously employed for boundary detection. Second, we incorporate 3D as well as 2D filters, to enable computations that use 3D context. Finally, we adopt a recursively trained architecture in which a first network generates a preliminary boundary map that is provided as input along with the original image to a second network that generates a final boundary map. Backpropagation training is accelerated by ZNN, a new implementation of 3D convolutional networks that uses multicore CPU parallelism for speed. Our hybrid 2D-3D architecture could be more generally applicable to other types of anisotropic 3D images, including video, and our recursive framework for any image labeling problem.

show abstract

Section: Introductionmentioning

confidence: 99%