Dual-feature Fusion Attention Network for Small Object Segmentation

Fei, Xin; Li, Xiaojie; Shi, Canghong; Ren, Hongping; Mumtaz, Imran; Guo, J.; Wu, Yu; Luo, Yong; Lv, Jiancheng; Wu, Xi

doi:10.1016/j.compbiomed.2023.106985

Cited by 8 publications

(3 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Notably, Attention-based approaches for small object segmentation have been investigated in the studies conducted by Fei et al [11], Sang et al [12], and Zhang et al [13]. This methodology holds the potential to enhance the accuracy of segmenting a greater number of eggs within the ovary.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for Automated Egg Maturation Prediction of Atlantic Salmon using Ultrasound Imaging

Yari,

Naeve,

Bergtun

et al. 2024

Preprint

View full text Add to dashboard Cite

The Atlantic salmon maturation process has been studied for decades to increase the quantity and quality of the production in farming facilities. An important topic in this context is salmon egg maturation process. Ultrasound imaging is considered an effective tool for monitoring the egg development stage of salmon, but manual inspection is time-consuming and highly dependent on operator experience. We propose a method for automated monitoring of the egg maturation stage in salmon using deep learning, providing complimentary decisions on egg morphology. A segmentation network was developed to solve the challenge of separating and measuring individual eggs in the ovary. The segmentation part was combined with a classification network to determine the maturation stage of the eggs. Our model was able to segment eggs and classify their development stage with over 88% accuracy, outperforming established methods designed for similar tasks. A real-time application was developed which provided an estimation of size and maturity stage while scanning. The egg state estimation showed potential for replacing manual evaluations and can enable fully automatic evaluation of maturation in Atlantic salmon.

show abstract

Section: Introductionmentioning

confidence: 99%

Deep Learning for Automated Egg Maturation Prediction of Atlantic Salmon using Ultrasound Imaging

Yari,

Naeve,

Bergtun

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…In this study, we propose a 2.5D‐based V‐Net network framework. The network replaces the original 3D convolution with a separable 3D (S3D) convolution 29 to enable the network to have both inter‐slice and intra‐slice features, and incorporate a dual‐branch feature fusion module (DFFM) 30 and an reverse attention context module (RACM) 30 to enhance the lung nodule features and edge features for more accurate edge segmentation results. In addition, we use a central pooling layer instead of the uniformly distributed maximum pooling layer of the original network, which is used to retain more information about the desired target and improve the feature extraction performance.…”

Section: Introductionmentioning

confidence: 99%

Dual‐branch feature fusion S3D V‐Net network for lung nodules segmentation

Xu,

Du,

Yin

2024

J Applied Clin Med Phys

View full text Add to dashboard Cite

BackgroundAccurate segmentation of lung nodules can help doctors get more accurate results and protocols in early lung cancer diagnosis and treatment planning, so that patients can be better detected and treated at an early stage, and the mortality rate of lung cancer can be reduced.PurposeCurrently, the improvement of lung nodule segmentation accuracy has been limited by his heterogeneous performance in the lungs, the imbalance between segmentation targets and background pixels, and other factors. We propose a new 2.5D lung nodule segmentation network model for lung nodule segmentation. This network model can well improve the extraction of edge information of lung nodules, and fuses intra‐slice and inter‐slice features, which makes good use of the three‐dimensional structural information of lung nodules and can more effectively improve the accuracy of lung nodule segmentation.MethodsOur approach is based on a typical encoding‐decoding network structure for improvement. The improved model captures the features of multiple nodules in both 3‐D and 2‐D CT images, complements the information of the segmentation target's features and enhances the texture features at the edges of the pulmonary nodules through the dual‐branch feature fusion module (DFFM) and the reverse attention context module (RACM), and employs central pooling instead of the maximal pooling operation, which is used to preserve the features around the target and to eliminate the edge‐irrelevant features, to further improve the performance of the segmentation of the pulmonary nodules.ResultsWe evaluated this method on a wide range of 1186 nodules from the LUNA16 dataset, and averaging the results of ten cross‐validated, the proposed method achieved the mean dice similarity coefficient (mDSC) of 84.57%, the mean overlapping error (mOE) of 18.73% and average processing of a case is about 2.07 s. Moreover, our results were compared with inter‐radiologist agreement on the LUNA16 dataset, and the average difference was 0.74%.ConclusionThe experimental results show that our method improves the accuracy of pulmonary nodules segmentation and also takes less time than more 3‐D segmentation methods in terms of time.

show abstract

“…Consequently, they propose the use of ensemble approaches, random data partitioning, and method-level approaches as alternatives to oversampling, as they lack erroneous assumptions. On the other hand, in [ 36 , 37 ], loss functions and other adjustments are used to improve the texture of the edges of small structures.…”

Section: Introductionmentioning

confidence: 99%

Improving Surgical Scene Semantic Segmentation through a Deep Learning Architecture with Attention to Class Imbalance

Urrea,

Garcia-Garcia,

Kern

2024

Biomedicines

View full text Add to dashboard Cite

This article addresses the semantic segmentation of laparoscopic surgery images, placing special emphasis on the segmentation of structures with a smaller number of observations. As a result of this study, adjustment parameters are proposed for deep neural network architectures, enabling a robust segmentation of all structures in the surgical scene. The U-Net architecture with five encoder–decoders (U-Net5ed), SegNet-VGG19, and DeepLabv3+ employing different backbones are implemented. Three main experiments are conducted, working with Rectified Linear Unit (ReLU), Gaussian Error Linear Unit (GELU), and Swish activation functions. The applied loss functions include Cross Entropy (CE), Focal Loss (FL), Tversky Loss (TL), Dice Loss (DiL), Cross Entropy Dice Loss (CEDL), and Cross Entropy Tversky Loss (CETL). The performance of Stochastic Gradient Descent with momentum (SGDM) and Adaptive Moment Estimation (Adam) optimizers is compared. It is qualitatively and quantitatively confirmed that DeepLabv3+ and U-Net5ed architectures yield the best results. The DeepLabv3+ architecture with the ResNet-50 backbone, Swish activation function, and CETL loss function reports a Mean Accuracy (MAcc) of 0.976 and Mean Intersection over Union (MIoU) of 0.977. The semantic segmentation of structures with a smaller number of observations, such as the hepatic vein, cystic duct, Liver Ligament, and blood, verifies that the obtained results are very competitive and promising compared to the consulted literature. The proposed selected parameters were validated in the YOLOv9 architecture, which showed an improvement in semantic segmentation compared to the results obtained with the original architecture.

show abstract

Dual-feature Fusion Attention Network for Small Object Segmentation

Cited by 8 publications

References 32 publications

Deep Learning for Automated Egg Maturation Prediction of Atlantic Salmon using Ultrasound Imaging

Deep Learning for Automated Egg Maturation Prediction of Atlantic Salmon using Ultrasound Imaging

Dual‐branch feature fusion S3D V‐Net network for lung nodules segmentation

Improving Surgical Scene Semantic Segmentation through a Deep Learning Architecture with Attention to Class Imbalance

Contact Info

Product

Resources

About