Multimodal Brain Tumor Segmentation Based on an Intelligent UNET-LSTM Algorithm in Smart Hospitals

Abstract

Smart hospitals are important components of smart cities. An intelligent medical system for brain tumor segmentation is required to construct smart hospitals. To achieve intelligent brain tumor segmentation, morphological variety and serious category imbalance must be managed effectively. Conventional deep neural networks have difficulty in predicting high-accuracy segmentation images due to these issues. To solve these problems, we propose using multimodal brain tumor images combined with the UNET and LSTM models to construct a new network structure with a mixed loss function to solve sample imbalance and describe an intelligent segmentation process to identify brain tumors. To verify the practicability of this algorithm, we used the open source Brain Tumor Segmentation Challenge dataset to train and verify the proposed network. We obtained DSCs of 0.91, 0.82, and 0.80; sensitivities of 0.93, 0.85, and 0.82; and specificities of 0.99, 0.99, and 0.98 in three tumor regions, including the whole tumor ( WT ), tumor core ( TC ), and enhanced tumor ( ET ). We also compared the results of the proposed network with those of other brain tumor segmentation methods, and the results showed that the proposed algorithm could segment different tumor lesions more accurately, highlighting its potential application value in the clinical diagnosis of brain tumors.

Brain tumor image segmentation algorithm based on multimodal feature fusion of Bayesian weight distribution

“…In comparison to the method proposed by Yu et al, 27 our approach utilizes an independent encoder to extract modal‐specific features, preserving the crucial characteristics of each modality without requiring any modal data preprocessing. On the other hand, when compared to the method presented by Hu et al, 30 our proposed method may exhibit slightly lower metrics in certain regions, but it outperforms in terms of overall brain tumor segmentation. These observations validate the effectiveness of our proposed method.…”

“…We compare the segmentation performance of this method with other multimodal brain tumor image segmentation methods on BraTs2018 and BraTs2019 datasets, as shown in Tables 3 and 4. The comparison method on the BraTS2018 dataset is introduced as follows: Zhou et al 26 proposed a correlation model specifically representing potential multi‐source correlations. Yu et al 27 proposed a sample adaptive intensity lookup table that dynamically converts the luminance contrast of each input MR image to accommodate subsequent segmentation tasks. Li et al 28 proposed a novel two‐stage cascaded network based on ResUNet, which is used to segment brain tumors and their substructures to solve the class imbalance problem in brain tumor data. Zhang et al 29 proposed a segmentation model that combines U‐Net with attention gates. Hu et al 30 proposed to use multimodal brain tumor images combined with UNET and LSTM models to construct a new network structure with mixed loss function to solve the problem of sample imbalance and realize the segmentation of brain tumor regions. …”

“…5. Hu et al 30 proposed to use multimodal brain tumor images combined with UNET and LSTM models to construct a new network structure with mixed loss function to solve the problem of sample imbalance and realize the segmentation of brain tumor regions.…”

Improving automatic segmentation of liver tumor images using a deep learning model

“…Therefore, they proposed a recalibration block that combines feature reorganization with spatial adaptability, offering a more suitable approach for semantic segmentation. Reference [ 32 ] introduces a novel network structure combining UNET and LSTM models with a mixed loss function, addressing morphological variety and category imbalance in brain tumor segmentation. Utilizing multimodal brain tumor images, the proposed method demonstrates superior accuracy in segmenting various tumor lesions, as evidenced by high DSCs, sensitivities, and specificities.…”

Brain tumour segmentation based on an improved U-Net

“…However, most of the current clinical brain tumour diagnoses are based on the clinician's experience. The method of manually segmenting, diagnosing and annotating tumour images is inefficient and demanding for image analysts, and it is easy to miss the best treatment window for patients [3]. Therefore, how to efficiently diagnose brain tumour images and reduce image diagnostic error has become a research direction for many researchers.…”