TAE-Seg: Generalized Lung Segmentation via Tilewise AutoEncoder Enhanced Network

Chen, Yurong; Zhang, Hui; Wang, Yaonan; Li-zhu, Liu; Wu, Q. M. Jonathan; Yang, Yimin

doi:10.1109/tim.2022.3217870

Cited by 5 publications

(6 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When using cross-datasets, our approach performed better than other approaches which used the SZ dataset for training and the MC dataset for validation. It is the case of studies described in [35] and [37], where the TAE-Seg and MS-AdaNet models reached 92.5% and 95.8% Dice scores respectively. In the same setting, our model performs better with 96.2%.…”

Section: Comparison To State-of-the-art Modelsmentioning

confidence: 96%

“…To face the above-mentioned challenges and obstacles, a diverse range of DL-based techniques for lung segmentation in CXR images has been studied in the literature in the last few years [31][32][33][34][35][36][37][38][39][40]. Newly, in 2023, MWG-UNet, a framework based on the Wasserstein generative adversarial network Ushape network to segment the lung fields and heart CXRs was presented [31].…”

Section: Related Workmentioning

confidence: 99%

“…This adjustment was made to extract multi-scale context features, each with different receptive field sizes. In [35], an unsupervised tilewise autoencoder (T-AE) pretraining architecture to acquire transferable knowledge was assessed, using an U-Net segmentation model for the fine-tuning process. In that year, HybridGNet, an UNet-based architecture was introduced [36].…”

Section: Related Workmentioning

confidence: 99%

“…Dice score. Also, in [35] and [37], the mentioned models were trained with images from the MC dataset and validated in the SZ dataset, reaching 95.5% and 94.1% Dice scores. Again, our model outperformed these results with a 95.6% Dice score using the same cross-datasets approach.…”

Section: Comparison To State-of-the-art Modelsmentioning

confidence: 99%

See 3 more Smart Citations

Automatic lung segmentation in chest X-ray images using SAM with prompts from YOLO

Khalili,

Priego-Torres,

Leon-Jimenez

et al. 2024

Preprint

View full text Add to dashboard Cite

Despite the impressive performance of current deep learning models in the field of medical imaging, the transfer of the lung segmentation task in X-ray images to clinical practice is still a pending task. In this study, we explored the performance of a fully automatic framework for lung fields segmentation in chest X-ray images, based on the combination of the Segment Anything Model (SAM) with prompt capabilities, and the You Only Look Once (YOLO) model to provide effective prompts. Transfer learning, loss functions and several validation strategies were intensively assessed. This provides a complete benchmark that enables future research studies to fairly compare new segmentation strategies. The results achieved demonstrate significant robustness and generalization capability against the variability in sensors, populations, disease manifestations, device processing and imaging conditions. The proposed framework is computationally efficient, can address bias in training over multiple datasets, and has the potential to be applied across other domains and modalities.

show abstract

Section: Comparison To State-of-the-art Modelsmentioning

confidence: 96%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Comparison To State-of-the-art Modelsmentioning

confidence: 99%

See 2 more Smart Citations

Automatic lung segmentation in chest X-ray images using SAM with prompts from YOLO

Khalili,

Priego-Torres,

Leon-Jimenez

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Thus, they brought more strength to the local connection between the adjacent chunks. In [25] Later came the concept of deeply learned vectors' formation [30], proposed by Naeem et al, mainly based on implementing a CNN with auto-correlation, gradient computation, scaling, filter, and localization coupled with state-of-the-art content-based image retrieval methods.…”

Section: B Literature Review 1) U-net and Its Variantsmentioning

confidence: 99%

EMED-UNet: An Efficient Multi-Encoder-Decoder Based UNet for Medical Image Segmentation

Shah,

Patel,

Thaker

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Many current and state-of-the-art deep learning models for accurate image segmentation are based on the U-Net architecture, a convolutional neural network designed for biomedical applications. Despite its widespread adoption in the medical imaging community, U-Net has two major limitations. First, due to its deep structure and the large number of filters used, the number of parameters and Floating-Point Operations Per Second (FLOPS) are high. This results in high computational complexity and demands a large memory size, making real-time implementation and deployment of U-Net models challenging. Second, the base U-Net model only uses a single type of kernel (i.e., 3×3) throughout the network for all convolution operations. Feature extraction using a single spatial extent throughout the network is suitable if the size, location, and shape of the salient regions remain static in all the images of a dataset, which is not necessarily the case in medical imaging. To address these two limitations, we propose an Efficient Multi-Encoder-Decoder based UNet (EMED-UNet), a novel architecture for efficient medical image segmentation. We evaluated our network on four medical imaging datasets: Montgomery County, Shenzhen CXR, COVID-19 CT LS, and BraTS (Brain Tumor Segmentation) dataset. EMED-UNet outperforms the U-Net and its variants in terms of accuracy, with around 77% reduction in parameters, a 60% reduction in FLOPS, and a 79.2% reduction in memory usage (all as compared to U-Net). The results demonstrate that EMED-UNet is a lightweight and accurate model for image segmentation that substantially improves upon the U-Net base model and is more feasible to deploy, given its decreased computational cost.

show abstract

A review on autoencoder based representation learning for fault detection and diagnosis in industrial processes

Qian

Song

Yao

et al. 2022

Chemometrics and Intelligent Laboratory Systems

View full text Add to dashboard Cite

TAE-Seg: Generalized Lung Segmentation via Tilewise AutoEncoder Enhanced Network

Cited by 5 publications

References 41 publications

Automatic lung segmentation in chest X-ray images using SAM with prompts from YOLO

Automatic lung segmentation in chest X-ray images using SAM with prompts from YOLO

EMED-UNet: An Efficient Multi-Encoder-Decoder Based UNet for Medical Image Segmentation

A review on autoencoder based representation learning for fault detection and diagnosis in industrial processes

Contact Info

Product

Resources

About