Nuclei segmentation using attention aware and adversarial networks

Neurocomputing, Journal Year: 2024, Volume and Issue: 579, P. 127445 - 127445

Published: Feb. 20, 2024

Language: Английский

---

ResDense-focal-DeepLabV3+ enabled litchi branch semantic segmentation for robotic harvesting

Computers and Electronics in Agriculture, Journal Year: 2023, Volume and Issue: 206, P. 107691 - 107691

Published: Feb. 15, 2023

Language: Английский

---

A systematic review of few-shot learning in medical imaging

Artificial Intelligence in Medicine, Journal Year: 2024, Volume and Issue: 156, P. 102949 - 102949

Published: Aug. 16, 2024

The lack of annotated medical images limits the performance deep learning models, which usually need large-scale labelled datasets. Few-shot techniques can reduce data scarcity issues and enhance image analysis speed robustness. This systematic review gives a comprehensive overview few-shot methods for analysis, aiming to establish standard methodological pipeline future research reference. With particular emphasis on role meta-learning, we analysed 80 relevant articles published from 2018 2023, conducting risk bias assessment extracting information, especially regarding employed techniques. From this, delineated shared among all studies. In addition, performed statistical studies' results concerning clinical task meta-learning method while also presenting supplemental information such as imaging modalities model robustness evaluation We discussed findings our providing insight into limitations state-of-the-art most promising approaches. Drawing investigation, yielded recommendations potential directions bridge gap between practice.

Language: Английский

---

Brain Stroke Lesion Segmentation Using Computed Tomography Images based on Modified U-Net Model with ResNet Blocks

International Journal of Online and Biomedical Engineering (iJOE), Journal Year: 2022, Volume and Issue: 18(13), P. 97 - 112

Published: Oct. 19, 2022

Segmentation of brain regions affected by ischemic stroke helps to overcome the main obstacles in modern studies visualization. Unfortunately, contemporary methods solving this problem using artificial intelligence are not optimal. Therefore, study we consider how increase efficiency segmentation focus computer perfusion imaging modifications based on UNet. The network was trained and tested ISLES 2018 dataset. publication includes an analysis results obtained, as well recommendations for future research. By choosing appropriate model parameters, our approach can be easily applied detect stroke. We present modified U-Net models with two ResNet blocks U-Net+ ResNetblock 1 +ResNetblock 2, a UNet model. Due small number images training model, best were obtained applying data preprocessing object representation approaches, normalization avoid overfitting. show that is superior other terms average distance recall, significant parameters

Language: Английский

---

Automated Multimodal Brain Tumor Segmentation and Localization in MRI Images Using Hybrid Res2-UNeXt

Journal of Electrical Engineering and Technology, Journal Year: 2024, Volume and Issue: 19(5), P. 3485 - 3497

Published: Feb. 3, 2024

Language: Английский

---

SPMUNet: Semantic segmentation of citrus surface defects driven by superpixel feature

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 224, P. 109182 - 109182

Published: June 26, 2024

Language: Английский

---

CA‐Unet++: An improved structure for medical CT scanning based on the Unet++ Architecture

International Journal of Intelligent Systems, Journal Year: 2022, Volume and Issue: 37(11), P. 8814 - 8832

Published: Aug. 10, 2022

Currently, deep learning has become more and mature in the field of medical image segmentation. Through using computer, models established can completely help doctors to perform Most current are based on Unet. The U-shaped structure skip connection layer Unet effectively achieve precise However, for complicated images, network is not sufficient enough. In response this problem, some scholars have designed Unet++ by adding a denser U-Net. Compared U-Net, effective dealing with complex but it drawbacks many aspects, there still large loss eigenvalues up-sampling processes. To address these issues, paper uses channel attention mechanism improve model obtain better segmentation efficiency accuracy. Meanwhile, Unet++, designs new called CA-Unet++. proposed module solve loses long-distance process process, respectively. experimental results data analysis shows that our CA-Unet++ performance computed tomography scan

Language: Английский

---

Asymmetric Cascade Fusion Network for Building Extraction

IEEE Transactions on Geoscience and Remote Sensing, Journal Year: 2023, Volume and Issue: 61, P. 1 - 18

Published: Jan. 1, 2023

The U-Net-like model has been widely studied in the field of building extraction. However, most these models are based on locally sensed Convolutional Neural Networks(CNNs) designed with symmetric structure and single feature processing, which cannot accurately identify buildings different sizes, shapes, colors remote sensing images. To overcome problems, we propose asymmetric cascade fusion network(ACFN), Vision Transformer(ViT), to design a novel architecture recognize sizes shapes by processing multi-granularity features means. First, obtains global contextual information embedding types attention encoder-decoders sizes. This can densely distributed occluded semantic reasoning images complex information. Second, multi-branch weighted pyramid pooling module, sets branch weights offset background noise introduced introducing Our ACFN significantly improves Beijing buildings, ISPRS-Vaihingen, LoveDA datasets.

Language: Английский

---

A survey on automated cell tracking: challenges and solutions

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: 83(34), P. 81511 - 81547

Published: March 9, 2024

Language: Английский

---

The 3D narrow butt weld seam detection system based on the binocular consistency correction

Journal of Intelligent Manufacturing, Journal Year: 2022, Volume and Issue: 34(5), P. 2321 - 2332

Published: March 17, 2022

Language: Английский

---