A5055757900- Cell Image Analysis Techniques
- Advanced Electron Microscopy Techniques and Applications
- Medical Image Segmentation Techniques
- Advanced Neuroimaging Techniques and Applications
- Electron and X-Ray Spectroscopy Techniques
- Functional Brain Connectivity Studies
- Image Processing Techniques and Applications
- Retinal Imaging and Analysis
- Genetics, Bioinformatics, and Biomedical Research
- Digital Imaging for Blood Diseases
- Neural dynamics and brain function
- Genomics and Phylogenetic Studies
- Force Microscopy Techniques and Applications
- Anomaly Detection Techniques and Applications
- Neural Networks and Applications
- Advancements in Photolithography Techniques
- Radiomics and Machine Learning in Medical Imaging
- Photoreceptor and optogenetics research
- Optical measurement and interference techniques
- Advanced Neural Network Applications
- Advanced X-ray Imaging Techniques
- Machine Learning in Materials Science
- Cellular Mechanics and Interactions
- Advanced Fluorescence Microscopy Techniques
Harvard University Press
2012-2018
Harvard University
2013-2017
ETH Zurich
2008-2012
State-of-the-art light and electron microscopes are capable of acquiring large image datasets, but quantitatively evaluating the data often involves manually annotating structures interest. This process is time-consuming a major bottleneck in evaluation pipeline. To overcome this problem, we have introduced Trainable Weka Segmentation (TWS), machine learning tool that leverages limited number manual annotations order to train classifier segment remaining automatically. In addition, TWS can...
In the field of neuroanatomy, automatic segmentation electron microscopy images is becoming one main limiting factors in getting new insights into functional structure brain. We propose a novel framework for thin elongated structures like membranes neuroanatomy setting. The probability output random forest classifier used regular cost function, which enforces gap completion via perceptual grouping constraints. global solution efficiently found by graph cut optimization. demonstrate...
Connectomics has recently begun to image brain tissue at nanometer resolution, which produces petabytes of data. This data must be aligned, labeled, proofread, and formed into graphs, each step this process requires visualization for human verification. As such, we present the BUTTERFLY middleware, a scalable platform that can handle massive interactive in connectomics. Our outputs geometry suitable hardware-accelerated rendering, abstracts low-level wrangling enable faster development new...
Automatic cell image segmentation methods in connectomics produce merge and split errors, which require correction through proofreading. Previous research has identified the visual search for these errors as bottleneck interactive To aid error correction, we develop two classifiers that automatically recommend candidate merges splits to user. These use a convolutional neural network (CNN) been trained with automatic segmentations against expert-labeled ground truth. Our detect...
Reconstructing a synaptic wiring diagram, or connectome, from electron microscopy (EM) images of brain tissue currently requires many hours manual annotation proofreading (Kasthuri and Lichtman, 2010; Lichtman Sanes, 2008; Seung, 2009). The desire to reconstruct ever larger more complex networks has pushed the collection EM datasets. A cubic millimeter raw imaging data would take up 1 PB storage present an project that be impractical without relying heavily on automatic segmentation methods....
Connectomics is the study of dense structure neurons in brain and their synapses, providing new insights into relation between brain's its function.Recent advances Electron Microscopy enable high-resolution imaging (4nm per pixel) neural tissue at a rate roughly 10 terapixels single day, allowing neuroscientists to capture large blocks reasonable amount time.The amounts data require novel computer vision based algorithms scalable software frameworks process this data.We describe RhoANA [1],...
An abstract is not available for this content so a preview has been provided. As you have access to content, full PDF via the 'Save PDF' action button.
Automatic, defect tolerant registration of transmission electron microscopy (TEM) images poses an important and challenging problem for biomedical image analysis, e.g. in computational neuroanatomy. In this paper we demonstrate a fully automatic stitching distortion correction method TEM propose probabilistic approach registration. The technique identifies defects due to sample preparation acquisition by outlier detection. A polynomial kernel expansion is used estimate non-linear...
We present an interactive approach to train a deep neural network pixel classifier for the segmentation of neuronal structures. An training scheme reduces extremely tedious manual annotation task that is typically required networks perform well on image problems. Our proposed method employs feedback loop captures sparse annotations using graphical user interface, trains based recent and past annotations, displays prediction output users in almost real-time. implementation algorithm also...
Automated sample preparation and electron microscopy enables acquisition of very large image data sets. These technical advances are special importance to the field neuroanatomy, as 3D reconstructions neuronal processes at nm scale can provide new insight into fine grained structure brain. Segmentation large-scale is main bottleneck in analysis these In this paper we present a pipeline that provides state-of-the art reconstruction performance while scaling sets GB-TB range. First, train...
Extract Extended abstract of a paper presented at Microscopy and Microanalysis 2007 in Ft. Lauderdale, Florida, USA, August 5 – 9,
We present an interactive approach to train a deep neural network pixel classifier for the segmentation of neuronal structures. An training scheme reduces extremely tedious manual annotation task that is typically required networks perform well on image problems. Our proposed method employs feedback loop captures sparse annotations using graphical user interface, trains based recent and past annotations, displays prediction output users in almost real-time. implementation algorithm also...
Connectomics is a field of neuroscience that analyzes neuronal connections. A connectome complete map system, comprising all connections between its structures. The term "connectome" close to the word "genome" and implies completeness connections, in same way as genome listing nucleotide sequences. goal connectomics create representation brain's wiring. Such believed increase our understanding how functional brain states emerge from their underlying anatomical structure. Furthermore, it can...
Automatic cell image segmentation methods in connectomics produce merge and split errors, which require correction through proofreading. Previous research has identified the visual search for these errors as bottleneck interactive To aid error correction, we develop two classifiers that automatically recommend candidate merges splits to user. These use a convolutional neural network (CNN) been trained with automatic segmentations against expert-labeled ground truth. Our detect...