Causal Imitation Learning-Based Navigation Algorithm for Drones

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 213 - 227

Published: Jan. 1, 2025

Language: Английский

---

LoC-SERS detection platform based on targeted signal anchoring mechanism, high-sensitivity detection of protein biomarkers in precancerous lesions of gastric cancer

Talanta, Journal Year: 2025, Volume and Issue: 294, P. 128190 - 128190

Published: April 18, 2025

Language: Английский

---

Enhanced medical image segmentation via dynamic and static attention aggregation

The Visual Computer, Journal Year: 2025, Volume and Issue: unknown

Published: May 8, 2025

Language: Английский

---

State-of-the-Art Deep Learning Methods for Microscopic Image Segmentation: Applications to Cells, Nuclei, and Tissues

Journal of Imaging, Journal Year: 2024, Volume and Issue: 10(12), P. 311 - 311

Published: Dec. 6, 2024

Microscopic image segmentation (MIS) is a fundamental task in medical imaging and biological research, essential for precise analysis of cellular structures tissues. Despite its importance, the process encounters significant challenges, including variability conditions, complex structures, artefacts (e.g., noise), which can compromise accuracy traditional methods. The emergence deep learning (DL) has catalyzed substantial advancements addressing these issues. This systematic literature review (SLR) provides comprehensive overview state-of-the-art DL methods developed over past six years microscopic images. We critically analyze key contributions, emphasizing how specifically tackle challenges cell, nucleus, tissue segmentation. Additionally, we evaluate datasets performance metrics employed studies. By synthesizing current identifying gaps existing approaches, this not only highlights transformative potential enhancing diagnostic research efficiency but also suggests directions future research. findings study have implications improving methodologies applications, ultimately fostering better patient outcomes advancing scientific understanding.

Language: Английский

---

DUCFNet: Dual U-shaped Cross-modal Fusion Network for Lung Infection Region Segmentation

Authorea (Authorea), Journal Year: 2024, Volume and Issue: unknown

Published: July 16, 2024

To promote further development of medical image segmentation, there is an increasing demand for high-quality datasets. Regrettably, are two major obstacles which the difficulty acquiring available images and financial burden data annotation constructing overcome difficulties, we leverage text to compensate defects existing In this work, propose a dual U-shaped network sufficiently achieve cross-modal feature fusion text. Specifically, one branches based on convolution neural network, named U-CNN, mainly extracts global features generate final prediction results. The other vision transformer blocks, U-ViT, responsible processing information merging from U-CNN. Additionally, utilize Cross-Attention Channel Fusion module Channel-wise Dual-branch Cross equip skip connection And modules greatly beneficial resolving semantic gaps enhancing integration information. Experimental results lung infection datasets with different modalities (X-Ray CT) suggest our method achieves excellent performance compared alternative state-of-the-art methods.

Language: Английский

---

Development and validation of a nomogram for obesity and related factors to detect gastric precancerous lesions in the Chinese population: a retrospective cohort study

Frontiers in Oncology, Journal Year: 2024, Volume and Issue: 14

Published: Nov. 20, 2024

Objectives The purpose of this study was to construct a nomogram identify patients at high risk gastric precancerous lesions (GPLs). This identification will facilitate early diagnosis and treatment ultimately reduce the incidence mortality cancer. Methods In single-center retrospective cohort study, 563 participants were divided into lesion (GPL) group (n=322) non-atrophic gastritis (NAG) (n=241) based on gastroscopy pathology results. Laboratory data demographic collected. A derivation (n=395) used factors associated with GPLs develop predictive model. Then, internal validation performed (n=168). We area under receiver operating characteristic curve (AUC) determine discriminative ability model; we constructed calibration plot evaluate accuracy decision analysis (DCA) assess clinical practicability Results Four –predictors (i.e., age, body mass index, smoking status, –triglycerides) included in AUC values model 0.715 (95% CI: 0.665-0.765) 0.717 0.640-0.795) cohorts, respectively. These indicated that had good discrimination ability. plots DCA suggested net benefit. Hosmer–Lemeshow test results cohorts for 0.774 0.468, Conclusion herein demonstrated performance terms predicting GPLs. can be beneficial detection GPLs, thus facilitating reducing

Language: Английский

---

MT-SCnet: multi-scale token divided and spatial-channel fusion transformer network for microscopic hyperspectral image segmentation

Frontiers in Oncology, Journal Year: 2024, Volume and Issue: 14

Published: Dec. 3, 2024

Hybrid architectures based on convolutional neural networks and Transformers, effectively captures both the local details overall structural context of lesion tissues cells, achieving highly competitive segmentation results in microscopic hyperspectral image (MHSI) tasks. However, fixed tokenization schemes single-dimensional feature extraction fusion existing methods lead to insufficient global pathology images.

Language: Английский

---

Large-scale chip layout pattern clustering method based on graph matching

Published: Dec. 10, 2024

In the integrated circuits field, rapid and accurate detection of defects anomalies is a critical factor in improving lithography process yields. Research on large-scale chip layout pattern feature extraction clustering algorithms plays crucial role enhancing manufacturing yield processes. This paper proposes graph matching-based method, leveraging high redundancy relatively simple circuit structure patterns. Our method innovatively employs graph-based representation to capture keypoint information patterns, applies dual-similarity constraints ensure both node edge similarities, utilizes agglomerative hierarchical merge structurally similar reducing reliance typical values. These enhancements allow for better handling complex geometries, thus efficiency stability clustering. Compared traditional methods based image statistical characteristics, our approach considers geometric within layout, achieving effective

Language: Английский

---

Photolithographic image prediction with conditional adversarial network and parameter encoding

Published: Dec. 10, 2024

Photolithography is a pivotal stage in integrated circuit chip manufacturing, exerting direct influence on both the performance and yield of chips. Its efficacy hinges heavily meticulous control parameters such as focus exposure dose. Traditionally, production speed limited by multiply rounds lengthy production-adjust process. Speeding up this process manufacturing has become pressing problem. To tackle challenge, we introduce novel framework that integrates conditional adversarial network (GAN) with parameter encoding module to predict SEM images from layout coupled photolithography parameters. During training phase, first pre-train model using paired data images, then fine-tune image corresponding lithography This proposed ensures generated are remarkably similar authentic images. Moreover, innovative structure allows GAN tailor generation according specific Extensive experiments validate effectiveness our method, indicating have constructed precise virtual capable predicting based inputs. approach not only effectively forecasts outcomes but also provides essential technical support address design challenges process, significantly streamlining path production.

Language: Английский

---