Mike Roberts - Academia.edu (original) (raw)

Papers by Mike Roberts

Automated sample preparation and electron microscopy enables acquisition of very large image data... more Automated sample preparation and electron microscopy enables acquisition of very large image data sets. These technical advances are of special importance to the field of neuroanatomy, as 3D reconstructions of neuronal processes at the nm scale can provide new insight into the fine grained structure of the brain. Segmentation of large-scale electron microscopy data is the main bottleneck in the analysis of these data sets. In this paper we present a pipeline that provides state-of-the art reconstruction performance while scaling to data sets in the GB-TB range. First, we train a random forest classifier on interactive sparse user annotations. The classifier output is combined with an anisotropic smoothing prior in a Conditional Random Field framework to generate multiple segmentation hypotheses per image. These segmentations are then combined into geometrically consistent 3D objects by segmentation fusion. We provide qualitative and quantitative evaluation of the automatic segmentation and demonstrate large-scale 3D reconstructions of neuronal processes from a 27,000μm(3) volume of brain tissue over a cube of 30μm in each dimension corresponding to 1000 consecutive image sections. We also introduce Mojo, a proofreading tool including semi-automated correction of merge errors based on sparse user scribbles.

IEEE Transactions on Visualization and Computer Graphics, 2014

Proofreading refers to the manual correction of automatic segmentations of image data. In connect... more Proofreading refers to the manual correction of automatic segmentations of image data. In connectomics, electron microscopy data is acquired at nanometer-scale resolution and results in very large image volumes of brain tissue that require fully automatic segmentation algorithms to identify cell boundaries. However, these algorithms require hundreds of corrections per cubic micron of tissue. Even though this task is time consuming, it is fairly easy for humans to perform corrections through splitting, merging, and adjusting segments during proofreading. In this paper we present the design and implementation of Mojo, a fully-featured single-user desktop application for proofreading, and Dojo, a multi-user web-based application for collaborative proofreading. We evaluate the accuracy and speed of Mojo, Dojo, and Raveler, a proofreading tool from Janelia Farm, through a quantitative user study. We designed a between-subjects experiment and asked non-experts to proofread neurons in a publicly available connectomics dataset. Our results show a significant improvement of corrections using web-based Dojo, when given the same amount of time. In addition, all participants using Dojo reported better usability. We discuss our findings and provide an analysis of requirements for designing visual proofreading software.

Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention, 2011

We present a novel semi-automatic method for segmenting neural processes in large, highly anisotr... more We present a novel semi-automatic method for segmenting neural processes in large, highly anisotropic EM (electron microscopy) image stacks. Our method takes advantage of sparse scribble annotations provided by the user to guide a 3D variational segmentation model, thereby allowing our method to globally optimally enforce 3D geometric constraints on the segmentation. Moreover, we leverage a novel algorithm for propagating segmentation constraints through the image stack via optimal volumetric pathways, thereby allowing our method to compute highly accurate 3D segmentations from very sparse user input. We evaluate our method by reconstructing 16 neural processes in a 1024 x 1024 x 50 nanometer-scale EM image stack of a mouse hippocampus. We demonstrate that, on average, our method is 68% more accurate than previous state-of-the-art semi-automatic methods.

Computer Methods and Programs in Biomedicine, 2013

3545; No. of Pages 8 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n ... more 3545; No. of Pages 8 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e x x x ( 2 0 1 3 ) xxx-xxx Segmentation Image processing Brain tumor Magnetic resonance imaging a b s t r a c t

The Biological Bulletin, 1997

In this paper, we introduce a novel framework for the compositing of interactively rendered 3D la... more In this paper, we introduce a novel framework for the compositing of interactively rendered 3D layers tailored to the needs of scientific illustration. Currently, traditional scientific illustrations are produced in a series of composition stages, combining different pictorial elements using 2D digital layering. Our approach extends the layer metaphor into 3D without giving up the advantages of 2D methods. The new compositing approach allows for effects such as selective transparency, occlusion overrides, and soft depth buffering. Furthermore, we show how common manipulation techniques such as masking can be integrated into this concept. These tools behave just like in 2D, but their influence extends beyond a single viewpoint. Since the presented approach makes no assumptions about the underlying rendering algorithms, layers can be generated based on polygonal geometry, volumetric data, point-based representations, or others. Our implementation exploits current graphics hardware and permits real-time interaction and rendering.

Explore a novel GPU level set segmentation algorithm that is both work-efficient and stepefficien... more Explore a novel GPU level set segmentation algorithm that is both work-efficient and stepefficient. Our algorithm has O(logn) step-complexity, in contrast to previous GPU algorithms which have O(n) step-complexity. We apply our algorithm to 3D medical images and we show that in typical clinical scenarios, our algorithm reduces the total number of processed level set field elements by 16x and is 14x faster than previous GPU algorithms with no reduction in segmentation accuracy.

Nutrition …, 2010

Background: This study's purpose investigated the impact of different macronutrient distributions... more Background: This study's purpose investigated the impact of different macronutrient distributions and varying caloric intakes along with regular exercise for metabolic and physiological changes related to weight loss.

We present a novel semi-automatic method for segmenting neural processes in large, highly anisotr... more We present a novel semi-automatic method for segmenting neural processes in large, highly anisotropic EM (electron microscopy) image stacks. Our method takes advantage of sparse scribble annotations provided by the user to guide a 3D variational segmentation model, thereby allowing our method to globally optimally enforce 3D geometric constraints on the segmentation. Moreover, we leverage a novel algorithm for propagating segmentation constraints through the image stack via optimal volumetric pathways, thereby allowing our method to compute highly accurate 3D segmentations from very sparse user input. We evaluate our method by reconstructing 16 neural processes in a 1024×1024×50 nanometer-scale EM image stack of a mouse hippocampus. We demonstrate that, on average, our method is 68% more accurate than previous state-of-the-art semi-automatic methods.

Deep Sea Research Part II: Topical Studies in Oceanography, 2014

The MESOBIO programme investigated mesoscale dynamics using an integrated ecosystem approach, lin... more The MESOBIO programme investigated mesoscale dynamics using an integrated ecosystem approach, linking physical and biogeochemical processes with different trophic levels. Observation and modeling were used in combination to explain the main processes occurring in the mesoscale eddy field. The particular shape of the Mozambique Channel, composed of two basins interconnected through a narrow zone, favours the generation of mesoscale eddies and increases the opportunity for eddy-shelf interactions. Phytoplankton abundance peaked in areas of nutrient enrichment that are often found in the core of cyclonic eddies, as well as on the continental shelf. Grazers in zooplankton communities exhibited high biovolume in cyclonic eddies, but their abundance was lower in fronts and divergence zones, with lowest biovolume in anticyclones. Biovolume was highest at shelf stations, but very variable and similar to phytoplankton. Age of eddies, their subsequent maturation stage and the dynamics of the eddy field played a major role effecting zooplankton abundance. Micronekton presented abundance patterns coherent with zooplankton distribution, however this was only demonstrated by acoustic methods, whereas mid-water trawl collection and predators stomach contents (predators being used as biological samplers) did not reveal significant relationships with mesoscale features. For upper trophic levels, the average density of foraging seabirds was lowest in anticyclones, highest in cyclones and at intermediate levels in divergence, shelf and frontal zones. However, multifaceted behavioral responses were observed in such a highly variable environment. Swordfish was clearly associated with divergence zones, and to a lesser extent with fronts, suggesting that the higher density in divergences was related to the presence of its main prey, essentially large squids. Although tunas tended to be more abundant in areas with weak geostrophic currents, their relationship to mesoscale features was not straightforward as adult tunas caught by longline have the ability to explore different foraging habitats over a broad range of depths. Several suggestions for advancing eddy-related research from the current state of knowledge are proposed in the second part of the paper.

Automated sample preparation and electron microscopy enables acquisition of very large image data... more Automated sample preparation and electron microscopy enables acquisition of very large image data sets. These technical advances are of special importance to the field of neuroanatomy, as 3D reconstructions of neuronal processes at the nm scale can provide new insight into the fine grained structure of the brain. Segmentation of large-scale electron microscopy data is the main bottleneck in the analysis of these data sets. In this paper we present a pipeline that provides state-of-the art reconstruction performance while scaling to data sets in the GB-TB range. First, we train a random forest classifier on interactive sparse user annotations. The classifier output is combined with an anisotropic smoothing prior in a Conditional Random Field framework to generate multiple segmentation hypotheses per image. These segmentations are then combined into geometrically consistent 3D objects by segmentation fusion. We provide qualitative and quantitative evaluation of the automatic segmentation and demonstrate large-scale 3D reconstructions of neuronal processes from a 27,000μm(3) volume of brain tissue over a cube of 30μm in each dimension corresponding to 1000 consecutive image sections. We also introduce Mojo, a proofreading tool including semi-automated correction of merge errors based on sparse user scribbles.

IEEE Transactions on Visualization and Computer Graphics, 2014

Proofreading refers to the manual correction of automatic segmentations of image data. In connect... more Proofreading refers to the manual correction of automatic segmentations of image data. In connectomics, electron microscopy data is acquired at nanometer-scale resolution and results in very large image volumes of brain tissue that require fully automatic segmentation algorithms to identify cell boundaries. However, these algorithms require hundreds of corrections per cubic micron of tissue. Even though this task is time consuming, it is fairly easy for humans to perform corrections through splitting, merging, and adjusting segments during proofreading. In this paper we present the design and implementation of Mojo, a fully-featured single-user desktop application for proofreading, and Dojo, a multi-user web-based application for collaborative proofreading. We evaluate the accuracy and speed of Mojo, Dojo, and Raveler, a proofreading tool from Janelia Farm, through a quantitative user study. We designed a between-subjects experiment and asked non-experts to proofread neurons in a publicly available connectomics dataset. Our results show a significant improvement of corrections using web-based Dojo, when given the same amount of time. In addition, all participants using Dojo reported better usability. We discuss our findings and provide an analysis of requirements for designing visual proofreading software.

Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention, 2011

We present a novel semi-automatic method for segmenting neural processes in large, highly anisotr... more We present a novel semi-automatic method for segmenting neural processes in large, highly anisotropic EM (electron microscopy) image stacks. Our method takes advantage of sparse scribble annotations provided by the user to guide a 3D variational segmentation model, thereby allowing our method to globally optimally enforce 3D geometric constraints on the segmentation. Moreover, we leverage a novel algorithm for propagating segmentation constraints through the image stack via optimal volumetric pathways, thereby allowing our method to compute highly accurate 3D segmentations from very sparse user input. We evaluate our method by reconstructing 16 neural processes in a 1024 x 1024 x 50 nanometer-scale EM image stack of a mouse hippocampus. We demonstrate that, on average, our method is 68% more accurate than previous state-of-the-art semi-automatic methods.

Computer Methods and Programs in Biomedicine, 2013

3545; No. of Pages 8 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n ... more 3545; No. of Pages 8 c o m p u t e r m e t h o d s a n d p r o g r a m s i n b i o m e d i c i n e x x x ( 2 0 1 3 ) xxx-xxx Segmentation Image processing Brain tumor Magnetic resonance imaging a b s t r a c t

The Biological Bulletin, 1997

In this paper, we introduce a novel framework for the compositing of interactively rendered 3D la... more In this paper, we introduce a novel framework for the compositing of interactively rendered 3D layers tailored to the needs of scientific illustration. Currently, traditional scientific illustrations are produced in a series of composition stages, combining different pictorial elements using 2D digital layering. Our approach extends the layer metaphor into 3D without giving up the advantages of 2D methods. The new compositing approach allows for effects such as selective transparency, occlusion overrides, and soft depth buffering. Furthermore, we show how common manipulation techniques such as masking can be integrated into this concept. These tools behave just like in 2D, but their influence extends beyond a single viewpoint. Since the presented approach makes no assumptions about the underlying rendering algorithms, layers can be generated based on polygonal geometry, volumetric data, point-based representations, or others. Our implementation exploits current graphics hardware and permits real-time interaction and rendering.

Explore a novel GPU level set segmentation algorithm that is both work-efficient and stepefficien... more Explore a novel GPU level set segmentation algorithm that is both work-efficient and stepefficient. Our algorithm has O(logn) step-complexity, in contrast to previous GPU algorithms which have O(n) step-complexity. We apply our algorithm to 3D medical images and we show that in typical clinical scenarios, our algorithm reduces the total number of processed level set field elements by 16x and is 14x faster than previous GPU algorithms with no reduction in segmentation accuracy.

Nutrition …, 2010

Background: This study's purpose investigated the impact of different macronutrient distributions... more Background: This study's purpose investigated the impact of different macronutrient distributions and varying caloric intakes along with regular exercise for metabolic and physiological changes related to weight loss.

We present a novel semi-automatic method for segmenting neural processes in large, highly anisotr... more We present a novel semi-automatic method for segmenting neural processes in large, highly anisotropic EM (electron microscopy) image stacks. Our method takes advantage of sparse scribble annotations provided by the user to guide a 3D variational segmentation model, thereby allowing our method to globally optimally enforce 3D geometric constraints on the segmentation. Moreover, we leverage a novel algorithm for propagating segmentation constraints through the image stack via optimal volumetric pathways, thereby allowing our method to compute highly accurate 3D segmentations from very sparse user input. We evaluate our method by reconstructing 16 neural processes in a 1024×1024×50 nanometer-scale EM image stack of a mouse hippocampus. We demonstrate that, on average, our method is 68% more accurate than previous state-of-the-art semi-automatic methods.

Deep Sea Research Part II: Topical Studies in Oceanography, 2014

The MESOBIO programme investigated mesoscale dynamics using an integrated ecosystem approach, lin... more The MESOBIO programme investigated mesoscale dynamics using an integrated ecosystem approach, linking physical and biogeochemical processes with different trophic levels. Observation and modeling were used in combination to explain the main processes occurring in the mesoscale eddy field. The particular shape of the Mozambique Channel, composed of two basins interconnected through a narrow zone, favours the generation of mesoscale eddies and increases the opportunity for eddy-shelf interactions. Phytoplankton abundance peaked in areas of nutrient enrichment that are often found in the core of cyclonic eddies, as well as on the continental shelf. Grazers in zooplankton communities exhibited high biovolume in cyclonic eddies, but their abundance was lower in fronts and divergence zones, with lowest biovolume in anticyclones. Biovolume was highest at shelf stations, but very variable and similar to phytoplankton. Age of eddies, their subsequent maturation stage and the dynamics of the eddy field played a major role effecting zooplankton abundance. Micronekton presented abundance patterns coherent with zooplankton distribution, however this was only demonstrated by acoustic methods, whereas mid-water trawl collection and predators stomach contents (predators being used as biological samplers) did not reveal significant relationships with mesoscale features. For upper trophic levels, the average density of foraging seabirds was lowest in anticyclones, highest in cyclones and at intermediate levels in divergence, shelf and frontal zones. However, multifaceted behavioral responses were observed in such a highly variable environment. Swordfish was clearly associated with divergence zones, and to a lesser extent with fronts, suggesting that the higher density in divergences was related to the presence of its main prey, essentially large squids. Although tunas tended to be more abundant in areas with weak geostrophic currents, their relationship to mesoscale features was not straightforward as adult tunas caught by longline have the ability to explore different foraging habitats over a broad range of depths. Several suggestions for advancing eddy-related research from the current state of knowledge are proposed in the second part of the paper.