摘要 :
The state-of-the-art models for medical image segmentation are variants of U-Net and fully convolutional networks (FCN). Despite their success, these models have two limitations: (1) their optimal depth is apriori unknown, requiri...
展开
The state-of-the-art models for medical image segmentation are variants of U-Net and fully convolutional networks (FCN). Despite their success, these models have two limitations: (1) their optimal depth is apriori unknown, requiring extensive architecture search or inefficient ensemble of models of varying depths; and (2) their skip connections impose an unnecessarily restrictive fusion scheme, forcing aggregation only at the same-scale feature maps of the encoder and decoder sub-networks. To overcome these two limitations, we propose UNet++, a new neural architecture for semantic and instance segmentation, by (1) alleviating the unknown network depth with an efficient ensemble of U-Nets of varying depths, which partially share an encoder and co-learn simultaneously using deep supervision; (2) redesigning skip connections to aggregate features of varying semantic scales at the decoder sub-networks, leading to a highly flexible feature fusion scheme; and (3) devising a pruning scheme to accelerate the inference speed of UNet++. We have evaluated UNet++ using six different medical image segmentation datasets, covering multiple imaging modalities such as computed tomography (CT), magnetic resonance imaging (MRI), and electron microscopy (EM), and demonstrating that (1) UNet++ consistently outperforms the baseline models for the task of semantic segmentation across different datasets and backbone architectures; (2) UNet++ enhances segmentation quality of varying-size objects-an improvement over the fixed-depth U-Net; (3) Mask RCNN++ (Mask R-CNN with UNet++ design) outperforms the original Mask R-CNN for the task of instance segmentation; and (4) pruned UNet++ models achieve significant speedup while showing only modest performance degradation. Our implementation and pre-trained models are available at https://github.com/MrGiovanni/UNetPlusPlus.
收起
摘要 :
Tremendous efforts have been made on image segmentation but the mask quality is still not satisfactory. The boundaries of predicted masks are usually imprecise due to the low spatial resolution of feature maps and the imbalance pr...
展开
Tremendous efforts have been made on image segmentation but the mask quality is still not satisfactory. The boundaries of predicted masks are usually imprecise due to the low spatial resolution of feature maps and the imbalance problem caused by the extremely low proportion of boundary pixels. To address these issues, we propose a conceptually simple yet effective post-processing refinement framework, termed BPR, to improve the boundary quality of the prediction of any image segmentation model. Following the idea of looking closer to segment boundaries better, we extract and refine a series of small boundary patches along the predicted boundaries. The refinement is accomplished by a boundary patch refinement network at the higher resolution. The trained BPR model can be easily transferred to refine the results of other models as well. Extensive experiments show that the proposed BPR framework yields significant improvements on the semantic, instance, and panoptic segmentation tasks over a variety of baselines on the Cityscapes dataset.
收起
摘要 :
This paper presents novel features and an architecture for an automatic on-line acoustic classification and segmentation system. The system includes speech/non-speech segmentation (with the emphasis on accurate speech/music segmen...
展开
This paper presents novel features and an architecture for an automatic on-line acoustic classification and segmentation system. The system includes speech/non-speech segmentation (with the emphasis on accurate speech/music segmentation), gender segmentation, and speech bandwidth segmentation. This automatic segmentation system can be easily integrated into an automatic continuous speech recognition system, where information about individual acoustic segments can be used for acoustic model selection and adaptation, or as additional information for rich transcription output. Acoustic model adaptation can improve the speech recognition rate and additional information in rich transcription can be useful when searching for some certain events or circumstances (male speaker talking over the phone line, etc.). For speech/non-speech segmentation we propose a new set of features, which are based on an energy variance in a narrow frequency sub-band, called EVFB (Energy Variance of Filter Bank). The proposed features also prove to be an efficient discriminator between speech and music. Segmentation cross-test results show that EVFB features prove to be more robust than MFCC features. Two new features (modified spectral roll-off and high-frequency energy variance) are also proposed for speech bandwidth classification and segmentation. The results show a good and robust performance by the automatic on-line acoustic segmentation system. All experiments and tests were performed on a radio broadcast database and a Slovenian BNSI Broadcast News database. Integration of the automatic on-line acoustic segmentation system into a continuous speech recognition system based on MFCC (mel-frequency cepstral coefficients) features requires only a small additional computational cost because many of the proposed system's feature calculation procedures are common to the MFCC features calculation procedure.
收起
摘要 :
Deep convolutional neural networks have shown outstanding performance in the task of semantically segmenting images. Applying the same methods on 3D data still poses challenges due to the heavy memory requirements and the lack of ...
展开
Deep convolutional neural networks have shown outstanding performance in the task of semantically segmenting images. Applying the same methods on 3D data still poses challenges due to the heavy memory requirements and the lack of structured data. Here, we propose LatticeNet, a novel approach for 3D semantic segmentation, which takes raw point clouds as input. A PointNet describes the local geometry which we embed into a sparse permutohedral lattice. The lattice allows for fast convolutions while keeping a low memory footprint. Further, we introduce DeformSlice, a novel learned data-dependent interpolation for projecting lattice features back onto the point cloud. We present results of 3D segmentation on multiple datasets where our method achieves state-of-the-art performance. We also extend and evaluate our network for instance and dynamic object segmentation.
收起
摘要 :
Due to the high diversity of image data, image segmentation is still a very challenging problem after decades of development. Each segmentation algorithm has its merits as well as its drawbacks. Instead of segmenting images via co...
展开
Due to the high diversity of image data, image segmentation is still a very challenging problem after decades of development. Each segmentation algorithm has its merits as well as its drawbacks. Instead of segmenting images via conventional techniques, inspired by the idea of the ensemble clustering technique that combines a set of weak clusterers to obtain a strong clusterer, we propose to achieve a consensus segmentation by fusing evidence accumulated from multiple weak segmentations (or over segmentations). We present a novel image segmentation approach which exploits multiple over segmentations and achieves segmentation results by hierarchical region merging. The cross-region evidence accumulation (CREA) mechanism is designed for collecting information among over segmentations. The pixel-pairs across regions are treated as a bag of independent voters and the cumulative votes from multiple over-segmentations are fused to estimate the coherency of adjacent regions. We further integrate the brightness, color, and texture cues for measuring the appearance similarity between regions in an over-segmentation, which, together with the CREA information, are utilized for making the region merging decisions. Experiments are conducted on multiple public data sets, which demonstrate the superiority of our-approach in terms of both effectiveness and efficiency when compared to the state-of-the-art. (C) 2016 Elsevier B.V. All rights reserved.
收起
摘要 :
Many medical image segmentation methods require the selection of seed points inside the target structure. Often times, the location of these seed points determines the accuracy of the resulting target structure delineation and may...
展开
Many medical image segmentation methods require the selection of seed points inside the target structure. Often times, the location of these seed points determines the accuracy of the resulting target structure delineation and may lead to undesirably high delineation variability. We present Robust-Seed, a new method for automatically reducing the variability of manual and semi-automatic seed-based segmentation methods with respect to the seed point location without compromising the target structure segmentation accuracy. The inputs are a volumetric image, a seed point inside the target structure, and a seed-based segmentation method. The output is a new seed point that optimises the target structure segmentation result. The algorithm iteratively computes a new seed point location that improves the expected target structure segmentation for the given method. Experimental evaluation of seed-based fast-marching level-set and adaptive region growing segmentation of the kidney and the liver on 32 CT scans with ground-truth delineations shows that Robust-Seed yields a perfect robustness score with no significant compromise on the segmentation quality (paired f-test,p < 0.05).The key advantages of Robust-Seed are that it is automatic, that it is independent of target structure and segmentation method used, and that it applies to a wide class of anatomical structures and clinical tasks.
收起
摘要 :
We present molecular-level simulations of dendrimer/DNA complexes in the presence of a model cell membrane. We determine the required conditions for the complex to arrive intact at the membrane, and the lifetime of the complex as ...
展开
We present molecular-level simulations of dendrimer/DNA complexes in the presence of a model cell membrane. We determine the required conditions for the complex to arrive intact at the membrane, and the lifetime of the complex as it resides attached to the membrane. Our simulations directly pertain to critical issues arising in emerging gene delivery therapeutic applications, where a molecular carrier is required to deliver DNA segments to the interior of living cells.
收起
摘要 :
Objectives: The main aim of the study was to examine the development and course of the facial nerve within fetal temporal bones from an anatomical and neuro-otological perspective.
摘要 :
? 2024 The Author(s)Despite being one of the most prevalent forms of cancer, prostate cancer (PCa) shows a significantly high survival rate, provided there is timely detection and treatment. Computational methods can help make thi...
展开
? 2024 The Author(s)Despite being one of the most prevalent forms of cancer, prostate cancer (PCa) shows a significantly high survival rate, provided there is timely detection and treatment. Computational methods can help make this detection process considerably faster and more robust. However, some modern machine-learning approaches require accurate segmentation of the prostate gland and the index lesion. Since performing manual segmentations is a very time-consuming task, and highly prone to inter-observer variability, there is a need to develop robust semi-automatic segmentation models. In this work, we leverage the large and highly diverse ProstateNet dataset, which includes 638 whole gland and 461 lesion segmentation masks, from 3 different scanner manufacturers provided by 14 institutions, in addition to other 3 independent public datasets, to train accurate and robust segmentation models for the whole prostate gland, zones and lesions. We show that models trained on large amounts of diverse data are better at generalizing to data from other institutions and obtained with other manufacturers, outperforming models trained on single-institution single-manufacturer datasets in all segmentation tasks. Furthermore, we show that lesion segmentation models trained on ProstateNet can be reliably used as lesion detection models.
收起
摘要 :
The methylmercury (MeHg) content in the Second Songhua River was investigated in this study. Compared with the former data, the following trends in temporal variation were observed. The MeHg content decreased in relation to the di...
展开
The methylmercury (MeHg) content in the Second Songhua River was investigated in this study. Compared with the former data, the following trends in temporal variation were observed. The MeHg content decreased in relation to the distance from the pollution source in 1977; it showed a decline in 1983 after the pollution source had been shut off for one year and continued to decline from 1983 to 1991 when the Second Songhua River was in a cleansing period. The MeHg content in most segments investigated in this study was higher than in 1991. Along the river, sediment samples were collected from the Baishan Segment to the Sifangtai Segment, and from the surface to the underlying layer to check the vertical trend. The MeHg content was high in the segments upstream due to the gold mines existing, with highest content in the Toudaogou Segment (10.0 ng/g). The MeHg content declined from the Jiapigou Segment (6.2 ng/g) to the Hongshi Segment (0.69 ng/g), and it gradually increased from the Hongshi Segment (0.69 ng/g) to the Hadawan Segment (1.8 ng/g). It was still lower than upstream. The MeHg content gradually decreased from the Hadawan Segment to the Zhaoyuan Segment. However, there was no clear trend from the Zhaoyuan Segment to the Sifangtai Segment, where the order of the MeHg content was the Sanzhan Segment the Sifangtai Segment the Zhaoyuan Segment the Laozhou Segment. The vertical variation in sediments showed that the MeHg content in the surface layer was higher than in the underlying layer in all segments with the exception of the Hongshi Segment and the Zhaoyuan Segment. The pollution index of MeHg content in the Second Songhua River was also discussed.
收起
原文获取