Filters








21,108 Hits in 5.2 sec

End-to-End Boundary Aware Networks for Medical Image Segmentation [article]

Ali Hatamizadeh, Demetri Terzopoulos, Andriy Myronenko
2019 arXiv   pre-print
Our experiments reveal that our approach yields more accurate segmentation results, which makes it promising for more extensive application to medical image segmentation.  ...  Our networks are designed to account for organ boundary information, both by providing a special network edge branch and edge-aware loss terms, and they are trainable end-to-end.  ...  For instance, a radiologist segmenting a liver from CT images would usually trace liver edges first, from which the internal segmentation mask is easily deduced.  ... 
arXiv:1908.08071v2 fatcat:7jp243qqfnccdjymdz6jtsqhgm

End-to-End Boundary Aware Networks for Medical Image Segmentation [article]

Ali Hatamizadeh, Demetri Terzopoulos, Andriy Myronenko
2019 bioRxiv   pre-print
Our experiments reveal that our approach yields more accurate segmentation results, which makes it promising for more extensive application to medical image segmentation.  ...  Our networks are designed to account for organ boundary information, both by providing a special network edge branch and edge-aware loss terms, and they are trainable end-to-end.  ...  For instance, a radiologist segmenting a liver from CT images would usually trace liver edges first, from which the internal segmentation mask is easily deduced.  ... 
doi:10.1101/770248 fatcat:odbyod6ozzdfphcnmb2k4gv33e

Associating Inter-image Salient Instances for Weakly Supervised Semantic Segmentation [chapter]

Ruochen Fan, Qibin Hou, Ming-Ming Cheng, Gang Yu, Ralph R. Martin, Shi-Min Hu
2018 Lecture Notes in Computer Science  
We also combine our method with Mask R-CNN for instance segmentation, and demonstrated for the first time the ability of weakly supervised instance segmentation using only keyword annotations.  ...  The proposed framework is general, and any state-of-the-art fully-supervised network structure can be incorporated to learn the segmentation network.  ...  China (Project Number 61521002, 61620106008, 61572264) and the Joint NSFC-ISF Research Program (project number 61561146393), the national youth talent support program, Tianjin Natural Science Foundation for  ... 
doi:10.1007/978-3-030-01240-3_23 fatcat:cl46tczynff2tdx2myeprmjvhq

Inf-Net: Automatic COVID-19 Lung Infection Segmentation from CT Images

Deng-Ping Fan, Tao Zhou, Ge-Peng Ji, Yi Zhou, Geng Chen, Huazhu Fu, Jianbing Shen, Ling Shao
2020 IEEE Transactions on Medical Imaging  
Then, the implicit reverse attention and explicit edge-attention are utilized to model the boundaries and enhance the representations.  ...  Moreover, to alleviate the shortage of labeled data, we present a semi-supervised segmentation framework based on a randomly selected propagation strategy, which only requires a few labeled images and  ...  [43] proposed an attention-based semi-supervised deep network for pelvic organ segmentation, in which a semi-supervised region-attention loss is developed to address the insufficient data issue for  ... 
doi:10.1109/tmi.2020.2996645 pmid:32730213 fatcat:227q3yiporecdjxeixcj4jemhe

Label-PEnet: Sequential Label Propagation and Enhancement Networks for Weakly Supervised Instance Segmentation [article]

Weifeng Ge, Sheng Guo, Weilin Huang, Matthew R. Scott
2020 arXiv   pre-print
Weakly-supervised instance segmentation aims to detect and segment object instances precisely, given imagelevel labels only.  ...  Unlike previous methods which are composed of multiple offline stages, we propose Sequential Label Propagation and Enhancement Networks (referred as Label-PEnet) that progressively transform image-level  ...  Similarly, we perform CRF [23] to obtain more accurate results of instance segmentation. Instance Segmentation Module.  ... 
arXiv:1910.02624v3 fatcat:c5w76g7ikrc6jgp5kzgpuh4lcu

Inf-Net: Automatic COVID-19 Lung Infection Segmentation from CT Images [article]

Deng-Ping Fan, Tao Zhou, Ge-Peng Ji, Yi Zhou, Geng Chen, Huazhu Fu, Jianbing Shen, Ling Shao
2020 arXiv   pre-print
Then, the implicit reverse attention and explicit edge-attention are utilized to model the boundaries and enhance the representations.  ...  Moreover, to alleviate the shortage of labeled data, we present a semi-supervised segmentation framework based on a randomly selected propagation strategy, which only requires a few labeled images and  ...  The two parts of L seg provide effective global (image-level) and local (pixel-level) supervision for accurate segmentation.  ... 
arXiv:2004.14133v4 fatcat:2mkxpdwi3bg3tdaqpifksucnie

Chicken Image Segmentation via Multi-Scale Attention-Based Deep Convolutional Neural Network

Wei Li, Yang Xiao, Xibin Song, Na Lv, Xinbo Jiang, Yan Huang, Jingliang Peng
2021 IEEE Access  
Accurate segmentation and analysis for each animal in surveillance video images will help poultry farmers to monitor and promote animal welfare.  ...  Then, we propose an effective end-to-end framework for chicken image segmentation, which can also be used for other animal image segmentation.  ...  network supervision with side-outputs.  ... 
doi:10.1109/access.2021.3074297 fatcat:b4gm7la7drfgjpol7ir5vr7mpu

End-to-End Instance Edge Detection [article]

Xueyan Zou, Haotian Liu, Yong Jae Lee
2022 arXiv   pre-print
Finally, we use a penalty reduced focal loss to effectively train the model with point supervision on instance edges, which can reduce annotation costs.  ...  We demonstrate highly competitive instance edge detection performance compared to state-of-the-art baselines, and also show that the proposed task and loss are complementary to instance segmentation and  ...  Feature Pyramid Network The final output of the transformer encoder is 1/32 of the original image, which is too low for edge detection.  ... 
arXiv:2204.02898v1 fatcat:stvqnilfojbozjb4s5g6nyiqky

Inf-Net: Automatic COVID-19 Lung Infection Segmentation from CT Scans [article]

Deng-Ping Fan, Tao Zhou, Ge-Peng Ji, Yi Zhou, Geng Chen, Huazhu Fu, Jianbing Shen, Ling Shao
2020 medRxiv   pre-print
Then, the implicit reverse attention and explicit edge-attention are utilized to model the boundaries and enhance the representations.  ...  Moreover, to alleviate the shortage of labeled data, we present a semi-supervised segmentation framework based on a randomly selected propagation strategy, which only requires a few labeled images and  ...  The two parts of L seg provide effective global (image-level) and local (pixel-level) supervision for accurate segmentation.  ... 
doi:10.1101/2020.04.22.20074948 fatcat:i633vxx3knanjp3u6cabdngdwy

Multi-evidence Filtering and Fusion for Multi-label Classification, Object Detection and Semantic Segmentation Based on Weakly Supervised Learning

Weifeng Ge, Sibei Yang, Yizhou Yu
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
When there exist image level labels only, it is challenging for weakly supervised algorithms to achieve accurate predictions.  ...  The entire process consists of four stages, including object localization in the training images, filtering and fusing object instances, pixel labeling for the training images, and task-specific network  ...  instance classification network and compute attention maps in this network to obtain more accurate pixel level class probabilities.  ... 
doi:10.1109/cvpr.2018.00139 dblp:conf/cvpr/GeYY18 fatcat:dkwcvetog5gatoya2aqwrucvoi

EAA-Net: Rethinking the Autoencoder Architecture with Intra-class Features for Medical Image Segmentation [article]

Shiqiang Ma, Xuejian Li, Jijun Tang, Fei Guo
2022 arXiv   pre-print
In this paper, we propose a light-weight end-to-end segmentation framework based on multi-task learning, termed Edge Attention autoencoder Network (EAA-Net), to improve edge segmentation ability.  ...  However, segmentation networks pay too much attention to the main visual difference between foreground and background, and ignores the detailed edge information, which leads to a reduction in the accuracy  ...  Inspired by the difference between semantic segmentation and instance segmentation, our method combines the intraclass and inter-class features to obtain accurate edge segmentation results. x i represents  ... 
arXiv:2208.09197v1 fatcat:q2hvbrnbijhdlaeebknynlbhsm

Multi-Evidence Filtering and Fusion for Multi-Label Classification, Object Detection and Semantic Segmentation Based on Weakly Supervised Learning [article]

Weifeng Ge, Sibei Yang, Yizhou Yu
2018 arXiv   pre-print
When there exist image level labels only, it is challenging for weakly supervised algorithms to achieve accurate predictions.  ...  The entire process consists of four stages, including object localization in the training images, filtering and fusing object instances, pixel labeling for the training images, and task-specific network  ...  for each class, we train a singlelabel object instance classification network and compute attention maps in this network to obtain more accurate pixel level class probabilities.  ... 
arXiv:1802.09129v1 fatcat:mgoa2e4ysncpbgnif7axsqrog4

Weakly-supervised Salient Instance Detection [article]

Xin Tian, Ke Xu, Xin Yang, Baocai Yin, Rynson W.H. Lau
2020 arXiv   pre-print
Inspired by this insight, we propose to use class and subitizing labels as weak supervision for the SID problem.  ...  Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization.  ...  It contains 500 images with instance-aware pixel labels for training. For our weakly-supervised training, we extract the numbers of salient instances of these images as our subitizing labels.  ... 
arXiv:2009.13898v1 fatcat:baighmhl5fafdbgm2b2aiasgqm

Semantic-Edge-Supervised Single-Stage Detector for Oriented Object Detection in Remote Sensing Imagery

Dujuan Cao, Changming Zhu, Xinxin Hu, Rigui Zhou
2022 Remote Sensing  
It helps the network pay attention to the edge features of targets at multiple scales to obtain more regression clues.  ...  In this paper, we innovatively combine semantic edge detection with arbitrary-oriented object detection and propose a feature enhancement network base on a semantic edge supervision module (SES) that realizes  ...  The instances in each image are annotated with horizontal and oriented bounding boxes containing 14,596 instances. There is no division standard for the original dataset.  ... 
doi:10.3390/rs14153637 fatcat:c2u4g2zsljcxzoo47q6s6t55sy

Weakly Supervised Instance Segmentation using Class Peak Response [article]

Yanzhao Zhou, Yi Zhu, Qixiang Ye, Qiang Qiu, Jianbin Jiao
2018 arXiv   pre-print
To the best of our knowledge, we for the first time report results for the challenging image-level supervised instance segmentation task.  ...  Weakly supervised instance segmentation with image-level labels, instead of expensive pixel-level masks, remains unexplored.  ...  Acknowledgements The authors are very grateful for support by the NSFC grant 61771447 / 61671427, BMSTC, and NSF.  ... 
arXiv:1804.00880v1 fatcat:fw6vkby7xze2jm7v6zija2wg74
« Previous Showing results 1 — 15 out of 21,108 results