In recent years, the diagnosis of brain tumors with the help of magnetic resonance imaging (MRI) methods has received significant attention. MRI techniques with substantial capabilities of displaying the internal structures of the human body have become one of the most widely used methods in this field. In the present study, a tumor is segmented after effectively preprocessing MRI images. Then, the main features are mined using a combination of the gray-level cooccurrence matrix and discrete wavelet transform. Finally, the mined features are fed into an optimized convolutional neural network (CNN)-based classification using a new improved metaheuristic technique, called balanced sparrow search algorithm (BSSA) for the final diagnosis to improve the efficiency of the CNN concerning consistency and accuracy. To verify the efficacy of the recommended algorithm, it is implemented on the whole brain atlas (WBA) database, and the results are compared with certain new and well-known methods. A comparative result also has been performed to the study, and the results show that the highest accuracy achieved by the recommended BSSA-CNN system is 93.65%. In addition, it is demonstrated that the specificity of 65.07% in the presented method yields results that are significantly better than those of the competing methods.

中文翻译：

---

基于深度学习和平衡麻雀搜索算法的最优脑肿瘤诊断

近年来，借助磁共振成像（MRI）方法对脑肿瘤的诊断受到了极大的关注。具有显示人体内部结构的强大能力的MRI技术已成为该领域使用最广泛的方法之一。在本研究中，肿瘤在有效预处理 MRI 图像后被分割。然后，使用灰度共生矩阵和离散小波变换的组合来挖掘主要特征。最后，使用一种新的改进的元启发式技术（称为平衡麻雀搜索算法（BSSA））将挖掘的特征输入到基于卷积神经网络（CNN）的优化分类中，用于最终诊断，以提高 CNN 在一致性和准确性方面的效率。为了验证推荐算法的有效性，在全脑图谱（WBA）数据库上实现，并将结果与​​某些新的和众所周知的方法进行比较。还对研究进行了比较结果，结果表明，推荐的 BSSA-CNN 系统达到的最高准确率为 93.65%。此外，还证明了所提出的方法中 65.07% 的特异性产生的结果明显优于竞争方法的结果。