Image Interpolation Using Multi-scale Attention-aware Inception Network

SAHIS-Net: a spectral attention and feature enhancement network for microscopic hyperspectral cholangiocarcinoma image segmentation

Cholangiocarcinoma (CCA) poses a significant clinical challenge due to its aggressive nature and poor prognosis. While traditional diagnosis relies on color-based histopathology, hyperspectral imaging (HSI) offers rich, high-dimensional data holding potential for more accurate diagnosis. However, extracting meaningful insights from this data remains challenging. This work investigates the application of deep learning for CCA segmentation in microscopic HSI images, and introduces two novel neural networks: (1) Histogram Matching U-Net (HM-UNet) for efficient image pre-processing, and (2) Spectral Attention based Hyperspectral Image Segmentation Net (SAHIS-Net) for CCA segmentation. SAHIS-Net integrates a novel Spectral Attention (SA) module for adaptively weighing spectral information, an improved attention-aware feature enhancement (AFE) mechanism for better providing the model with more discriminative features, and a multi-loss training strategy for effective early stage feature extraction. We compare SAHIS-Net against several general and CCA-specific models, demonstrating its superior performance in segmenting CCA regions. These results highlight the potential of our approach for segmenting medical HSI images.

MAEF-Net: MLP Attention for Feature Enhancement in U-Net based Medical Image Segmentation Networks

Medical image segmentation plays an important role in diagnosis. Since the introduction of U-Net, numerous advancements have been implemented to enhance its performance and expand its applicability. The advent of Transformers in computer vision has led to the integration of self-attention mechanisms into U-Net, resulting in significant breakthroughs. However, the inherent complexity of Transformers renders these networks computationally demanding and parameter-heavy. Recent studies have demonstrated that multilayer perceptrons (MLPs), with their simpler architecture, can achieve comparable performance to Transformers in natural language processing and computer vision tasks. Building upon these findings, we have enhanced the previously proposed "Enhanced-Feature-Four-Fold-Net" (EF 3-Net) by introducing an MLP-attention block to learn long-range dependencies and expand the receptive field. This enhanced network is termed "MLP-Attention Enhanced-Feature-four-fold-Net", abbreviated as "MAEF-Net". To further enhance accuracy while reducing computational complexity, the proposed network incorporates additional efficient design elements. MAEF-Net was evaluated against several general and specialized medical image segmentation networks using four challenging medical image datasets. The results demonstrate that the proposed network exhibits high computational efficiency and comparable or superior performance to EF 3-Net and several state-of-the-art methods, particularly in segmenting blurry objects.

Study on the value of multi-dimensional conformal radiotherapy and functional imaging in tumor bioimaging

BACKGROUND

To explore the diagnostic and therapeutic effects of multi-dimensional conformal radiotherapy (MD-CRT) combined with functional imaging in tumor bioimaging.

METHODS

A total of 150 cases of patients with brain metastases in First Affiliated Hospital of Xi'an Jiaotong University between December 2020 and December 2021 were retrospectively selected as the research cohort. Participants underwent whole brain imaging guided by MD-CRT combined with volume assessment to assess the patient's recent treatment effect, statistically acceptable dose irradiation treatment, and the incidence of adverse reactions. All patients were followed up to evaluate the long-term efficacy.

RESULTS

Among the 150 patients, 24 cases were in complete remission, 72 cases were in partial remission, 36 cases were in a stable condition, 18 cases were in deterioration, and the treatment of 96 cases (64.00%) was deemed effective. All participants were followed up, the mean survival was (62.37±1.24) months, 96 cases (64.00%) survived, and 54 cases (36.00%) had died. The average dose was (62.09±3.94) Gy. In terms of adverse reactions: brain edema occurred in 57 patients, accounting for 38.00%.

CONCLUSIONS

The MD-CRT and functional imaging techniques for patients with metastatic tumor have high therapeutic accuracy, are associated with improved local control rate, prolonged survival, and cause little damage to normal tissues, with significant therapeutic effect, and can be widely used in clinical practice.