MF2-Net: A multipath feature fusion network for medical image segmentation

Yamanakkanavar, Nagaraj; Lee, Bumshik

doi:10.1016/j.engappai.2022.105004

Cited by 22 publications

(4 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experimental results show that the CM-SegNet model outperforms other methods with better segmentation performance and shorter training time, which has high clinical application value. In another piece of work, Yamanakkanavar and Lee [25] introduced MF2-Net, a multipath feature fusion CNN encompassing multiple encoder paths to seize layer-specific multiscale information. Each encoder path employs a stacked asymmetric kernel module termed SGC to efficiently encode contextual particulars in high-level features and accurately conflate neighboring feature cues.…”

Section: Related Workmentioning

confidence: 99%

HMSAM-UNet: A Hierarchical Multi-Scale Attention Module-Based Convolutional Neural Network for Improved CT Image Segmentation

Liu,

Lu,

Lian

et al. 2024

IEEE Access

View full text Add to dashboard Cite

While modern deep learning methods have made significant progress in medical image segmentation, some challenges remain, including accurately capturing features at multiple scales, limited ability to detect critical regions, and susceptibility to noise and background interference. To address these challenges, a new neural network called HMSAM-UNet is introduced in this work. A novel module, the Hierarchical Multi-Scale Attention Module (HMSAM), was designed in HMSAM-UNet to improve the precision and accuracy of CT image segmentation significantly. Specifically, HMSAM integrates the Hierarchical Attention Mechanism and Inception Module via residual connections. The Hierarchical Attention Mechanism can highlight important regions by learning attention weights, dramatically enhancing the model's ability to perceive critical areas for more accurate localization and segmentation of target structures in CT images. Meanwhile, incorporating the Inception module effectively strengthens the network's capacity to capture multi-scale features, substantially improving the model's ability to comprehend the structural characteristics of CT images. The results show that the average loss achieved by the proposed model has a 50.04% reduction compared to the original U-Net architecture. Furthermore, compared to other deep learning models such as FCN, DeepLabV3, PSPNet, Unet, UNet++, and SegNet, the model proposed in this work attains an average Dice coefficient of 98.72, and an average IoU score of 97.46 on the three datasets, both of which are the highest among all compared models.

show abstract

Section: Related Workmentioning

confidence: 99%

HMSAM-UNet: A Hierarchical Multi-Scale Attention Module-Based Convolutional Neural Network for Improved CT Image Segmentation

Liu,

Lu,

Lian

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Deep learning models have transformed the segmentation of robotic instruments, allowing robotic systems to distinguish between different surgical tools and monitor their locations during surgery. Recently, Mahmood et al presented a calibrated network for instrument segmentation in surgery environments to incorporate outcomes of dual streams of dilated convolutions [5]. They still acknowledge and consider the need for fine-tuning the foundation model to reduce the error set.…”

Section: Introductionmentioning

confidence: 99%

RSSNet: A Fine-tuned Deep Learning Network for Robotic Surgical-tool Segmentation

Janjua,

Ali,

Farooq

2023

Preprint

View full text Add to dashboard Cite

<p>Robotic systems have significantly transformed modern surgical practice by improving surgical precision and minimizing invasiveness. Crucial to these systems is the accurate segmentation of surgical tools, which remains a challenging task due to the complex surgical scenarios and the diversity of tools used. This paper presents a novel deep learning architecture, the Robotic Surgical Tool Segmentation Network (RSSNet), designed to enhance the accuracy of surgical tool segmentation in robotic surgeries. The proposed RSSNet model combines the power of Atrous Spatial Pyramid Pooling (ASPP) with the efficiency of average pooling layers to extract and encode intricate details for accurate segmentation. A fine-tuning strategy on the average pooling layers further enhances the model’s performance, particularly for small and intricate details. The model was rigorously tested on two primary datasets, Kvasir and EndoVis 2017, along with additional surgical datasets, demonstrating robustness and versatility. In particular, RSSNet outperformed state-of-the-art models in all comparisons. An ablation study further confirmed the effectiveness of each component in the model architecture, particularly the fine-tuned average pooling layers and the ASPP module. The study offers a novel, efficient, and high-performing solution for surgical tool segmentation, an essential step toward fully autonomous robotic surgeries. With a performance score of 96.78%, this work provides a baseline for optimizing fine-tuning strategies and incorporating additional data modalities for enhanced performance.</p>

show abstract

“…The proposed model was then trained from the entire input slice of the MRI scan using uniform patches of the same size. As proven in our earlier study [ 27 , 28 ], the partition of uniform patches facilitates brain MRI localization by focusing on fine details within each patch. Based on multiple evaluation metrics, the proposed approach demonstrates significant improvements over recently developed brain MRI segmentation methods.…”

Section: Introductionmentioning

confidence: 99%

SM-SegNet: A Lightweight Squeeze M-SegNet for Tissue Segmentation in Brain MRI Scans

Yamanakkanavar

Choi

Lee

2022

Sensors

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel squeeze M-SegNet (SM-SegNet) architecture featuring a fire module to perform accurate as well as fast segmentation of the brain on magnetic resonance imaging (MRI) scans. The proposed model utilizes uniform input patches, combined-connections, long skip connections, and squeeze–expand convolutional layers from the fire module to segment brain MRI data. The proposed SM-SegNet architecture involves a multi-scale deep network on the encoder side and deep supervision on the decoder side, which uses combined-connections (skip connections and pooling indices) from the encoder to the decoder layer. The multi-scale side input layers support the deep network layers’ extraction of discriminative feature information, and the decoder side provides deep supervision to reduce the gradient problem. By using combined-connections, extracted features can be transferred from the encoder to the decoder resulting in recovering spatial information, which makes the model converge faster. Long skip connections were used to stabilize the gradient updates in the network. Owing to the adoption of the fire module, the proposed model was significantly faster to train and offered a more efficient memory usage with 83% fewer parameters than previously developed methods, owing to the adoption of the fire module. The proposed method was evaluated using the open-access series of imaging studies (OASIS) and the internet brain segmentation registry (IBSR) datasets. The experimental results demonstrate that the proposed SM-SegNet architecture achieves segmentation accuracies of 95% for cerebrospinal fluid, 95% for gray matter, and 96% for white matter, which outperforms the existing methods in both subjective and objective metrics in brain MRI segmentation.

show abstract

MF2-Net: A multipath feature fusion network for medical image segmentation

Cited by 22 publications

References 41 publications

HMSAM-UNet: A Hierarchical Multi-Scale Attention Module-Based Convolutional Neural Network for Improved CT Image Segmentation

HMSAM-UNet: A Hierarchical Multi-Scale Attention Module-Based Convolutional Neural Network for Improved CT Image Segmentation

RSSNet: A Fine-tuned Deep Learning Network for Robotic Surgical-tool Segmentation

SM-SegNet: A Lightweight Squeeze M-SegNet for Tissue Segmentation in Brain MRI Scans

Contact Info

Product

Resources

About