Selective Feature Aggregation Network with Area-Boundary Constraints for Polyp Segmentation

Fang, Yuqi; Chen, Cheng; Yuan, Yixuan; Tong, Kai-Yu

doi:10.1007/978-3-030-32239-7_34

Cited by 221 publications

(113 citation statements)

References 11 publications

Supporting

Mentioning

112

Contrasting

Order By: Relevance

“…The previous study similar to ours is the selective feature‐aggregation network with area‐boundary constraints for polyp segmentation developed by Fang et al 36 . The main difference in this study is that we use the dynamic task‐balancing strategy to improve the efficiency of multitask learning.…”

Section: Discussionmentioning

confidence: 98%

“…Our goal was to minimize the weighted sum of training losses of all tasks as.

\underset{normalΘ}{italicmin} {false\sum}_{i = 1}^{m} w_{i} L ()(, D_{i} ; normalΘ)

where

w_{i}

was the weight of the

i

th task. For the segmentation task, the optimization objective was to minimize the dice loss function

L_{Dice}

32 and categorical cross‐entropy loss

L_{CE}

36 as

L (D_{1} ; Θ) = L_{italicDice} + L_{italicCE}

…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Automatic segmentation of hippocampus in hippocampal sparing whole brain radiotherapy: A multitask edge‐aware learning

Qiu

Yang

et al. 2021

Medical Physics

View full text Add to dashboard Cite

This study aimed to improve the accuracy of the hippocampus segmentation through multitask edge-aware learning. Method: We developed a multitask framework for computerized hippocampus segmentation. We used three-dimensional (3D) U-net as our backbone model with two training objectives: (a) to minimize the difference between the targeted binary mask and the model prediction; and (b) to optimize an auxiliary edge-prediction task which is designed to guide the model detection of the weak boundary of the hippocampus in model optimization. To balance the multiple task objectives, we proposed an improved gradient normalization by adaptively adjusting the weight of losses from different tasks. A total of 247 T1-weighted MRIs including 131 without contrast and 116 with contrast were collected from 247 patients to train and validate the proposed method. Segmentation was quantitatively evaluated with the dice coefficient (Dice), Hausdorff distance (HD), and average Hausdorff distance (AVD). The 3D U-net was used for baseline comparison. We used a Wilcoxon signed-rank test to compare repeated measurements (Dice, HD, and AVD) by different segmentations. Results: Through fivefold cross-validation, our multitask edge-aware learning achieved Dice of 0.8483 AE 0.0036, HD of 7.5706 AE 1.2330 mm, and AVD of 0.1522 AE 0.0165 mm, respectively. Conversely, the baseline results were 0.8340 AE 0.0072, 10.4631 AE 2.3736 mm, and 0.1884 AE 0.0286 mm, respectively. With a Wilcoxon signed-rank test, we found that the differences between our method and the baseline were statistically significant (P < 0.05). Conclusion:Our results demonstrated the efficiency of multitask edge-aware learning in hippocampus segmentation for hippocampal sparing whole-brain radiotherapy. The proposed framework may also be useful for other low-contrast small organ segmentations on medical imaging modalities.

show abstract

Section: Discussionmentioning

confidence: 98%

“…Our goal was to minimize the weighted sum of training losses of all tasks as.

\underset{normalΘ}{italicmin} {false\sum}_{i = 1}^{m} w_{i} L ()(, D_{i} ; normalΘ)

where

w_{i}

was the weight of the

i

th task. For the segmentation task, the optimization objective was to minimize the dice loss function

L_{Dice}

32 and categorical cross‐entropy loss

L_{CE}

36 as

L (D_{1} ; Θ) = L_{italicDice} + L_{italicCE}

…”

Section: Methodsmentioning

confidence: 99%

Automatic segmentation of hippocampus in hippocampal sparing whole brain radiotherapy: A multitask edge‐aware learning

Qiu

Yang

et al. 2021

Medical Physics

View full text Add to dashboard Cite

show abstract

“…This map is then refined by a series of recurrent reverse attention layers. Fang et al [17] used a network with one encoder and two mutually constrained decoders, one for predicting areas and another for predicting boundaries. The network then aggregates the features.…”

Section: ) Biomedical Image Segmentationmentioning

confidence: 99%

Training on Polar Image Transformations Improves Biomedical Image Segmentation

et al. 2021

View full text Add to dashboard Cite

A key step in medical image-based diagnosis is image segmentation. A common use case for medical image segmentation is the identification of single structures of an elliptical shape. Most organs like the heart and kidneys fall into this category, as well as skin lesions, polyps, and other types of abnormalities. Neural networks have dramatically improved medical image segmentation results, but still require large amounts of training data and long training times to converge. In this paper, we propose a general way to improve neural network segmentation performance and data efficiency on medical imaging segmentation tasks where the goal is to segment a single roughly elliptically distributed object. We propose training a neural network on polar transformations of the original dataset, such that the polar origin for the transformation is the center point of the object. This results in a reduction of dimensionality as well as a separation of segmentation and localization tasks, allowing the network to more easily converge. Additionally, we propose two different approaches to obtaining an optimal polar origin: (1) estimation via a segmentation trained on non-polar images and (2) estimation via a model trained to predict the optimal origin. We evaluate our method on the tasks of liver, polyp, skin lesion, and epicardial adipose tissue segmentation. We show that our method produces state-of-the-art results for lesion, liver, and polyp segmentation and performs better than most common neural network architectures for biomedical image segmentation. Additionally, when used as a pre-processing step, our method generally improves data efficiency across datasets and neural network architectures.

show abstract

“…Furthermore, only the methods published in the last two years, which clearly describe the experimental arrangement for selection of the training and test data subsets, have been listed. Interested readers can find more about other recently proposed methods in [60][61][62]. The information about other related colonoscopy image analysis problems, use of handcrafted features, and use of different modalities can be found in [9,24,27,63,64].…”

Section: Comparative Analysismentioning

confidence: 99%

Polyp Segmentation with Fully Convolutional Deep Neural Networks—Extended Evaluation Study

2020

View full text Add to dashboard Cite

Analysis of colonoscopy images plays a significant role in early detection of colorectal cancer. Automated tissue segmentation can be useful for two of the most relevant clinical target applications—lesion detection and classification, thereby providing important means to make both processes more accurate and robust. To automate video colonoscopy analysis, computer vision and machine learning methods have been utilized and shown to enhance polyp detectability and segmentation objectivity. This paper describes a polyp segmentation algorithm, developed based on fully convolutional network models, that was originally developed for the Endoscopic Vision Gastrointestinal Image Analysis (GIANA) polyp segmentation challenges. The key contribution of the paper is an extended evaluation of the proposed architecture, by comparing it against established image segmentation benchmarks utilizing several metrics with cross-validation on the GIANA training dataset. Different experiments are described, including examination of various network configurations, values of design parameters, data augmentation approaches, and polyp characteristics. The reported results demonstrate the significance of the data augmentation, and careful selection of the method’s design parameters. The proposed method delivers state-of-the-art results with near real-time performance. The described solution was instrumental in securing the top spot for the polyp segmentation sub-challenge at the 2017 GIANA challenge and second place for the standard image resolution segmentation task at the 2018 GIANA challenge.

show abstract

Selective Feature Aggregation Network with Area-Boundary Constraints for Polyp Segmentation

Cited by 221 publications

References 11 publications

Automatic segmentation of hippocampus in hippocampal sparing whole brain radiotherapy: A multitask edge‐aware learning

Automatic segmentation of hippocampus in hippocampal sparing whole brain radiotherapy: A multitask edge‐aware learning

Training on Polar Image Transformations Improves Biomedical Image Segmentation

Polyp Segmentation with Fully Convolutional Deep Neural Networks—Extended Evaluation Study

Contact Info

Product

Resources

About