Search

Image contrast enhancement is a crucial computer vision step aiming to improve the quality of the visual information in processed images. In the literature, several proposed methods for image contrast enhancement are Histogram Equalization-based (HE) techniques that use one transformation function and optimize its parameters for mapping the pixels to new gray-intensity values. However, using only one transformation function would leave other enhancement options unexplored. Therefore, the proposed approach generates several transformation functions and selects the one that best improves the image's contrast. This method is based on the Differential Evolution (DE) algorithm, which produces multiple candidate solutions representing transformation functions. The transformation functions map the input pixel values in their enhanced versions to equalize the histogram and improve the image's contrast. Furthermore, a new formulation is proposed as the objective function based on the number of edge pixels, the intensity of the pixels, image entropy, and the number of gray intensity levels. The performance of this approach has been tested on low-contrast dataset images and compared to similar HE techniques, such as AVHEQ, BBHE, RSESIHE, MMBEBHE, and ESIHE. The results demonstrate the proposed algorithm's robustness and high performance in improving the grayscale images' contrast. [ABSTRACT FROM AUTHOR]

Hybridization of metaheuristic algorithms has recently been introduced to increase the search capabilities of a traditional metaheuristic algorithms. The fuzzy optimization approach (FOA) is a metaheuristic algorithm that implements fuzzy logic systems to incorporate the knowledge and experience of an expert metaheuristic designer. This integration enables the FOA to guide the search strategy throughout the optimization process. It has been widely applied in high-dimensional problems due to its robustness. However, it presents some issues, such as a slow convergence rate, low exploration and exploitation mechanisms, and high computational effort. In this paper, hybrid search mechanisms are implemented to the original structure of FOA to increase its performance in terms of search capabilities over the limits of the fitness functions. The proposed method called ODM-FOA combines the advantages of Metropolis Hasting (MH) initialization, opposition-based learning (OBL), and Diversity measures to correctly identify and register prominent areas within the search space by approximating the fitness function and improving the exploration and exploitation of the search space. A comparison between ODM-FOA and the original FOA is implemented. Additionally, to evaluate the performance of the developed scheme, it is compared among six well-known metaheuristic algorithms on a set of benchmark functions. Experimental results demonstrate the effectiveness and robustness of the proposed approach against the other original FOA and the other methodologies in terms of solution quality, dimensionality, similarity, and convergence criteria. [ABSTRACT FROM AUTHOR]

Most metaheuristic methods employ a strategy designed to be generical and fixed. As a result, these methods often lack the capability to adaptively modify their performance in response to different scenarios or challenges encountered during the search process. This paper presents a new metaheuristic algorithm designed to dynamically adjust its search strategy throughout the optimization process for increased efficiency. This algorithm is based on Evolutionary Strategies (ES) due to their notable self-adaptive features. To further enhance its efficiency, we have incorporated elements from Evolutionary Game Theory (EGT). This integration ensures a more comprehensive strategy adaptation process, taking into account not only the information from the specific agent but also insights from other population members. Additionally, our approach alters the conventional EGT mechanism by including not just pairwise evaluations but also data from the top-performing individuals in the population, based on their outcomes. This broader adaptation strategy allows for a faster convergence to the most effective dominant strategy. To demonstrate the effectiveness of our method, we compared it against several established metaheuristic algorithms using 28 diverse test functions. Our findings reveal that this approach produces competitive results, delivering higher-quality solutions and faster convergence rates. [ABSTRACT FROM AUTHOR]

Breast cancer remains a critical global concern, underscoring the urgent need for early detection and accurate diagnosis to improve survival rates among women. Recent developments in deep learning have shown promising potential for computer-aided detection (CAD) systems to address this challenge. In this study, a novel segmentation method based on deep learning is designed to detect tumors in breast ultrasound images. Our proposed approach combines two powerful attention mechanisms: the novel Positional Convolutional Block Attention Module (PCBAM) and Shifted Window Attention (SWA), integrated into a Residual U-Net model. The PCBAM enhances the Convolutional Block Attention Module (CBAM) by incorporating the Positional Attention Module (PAM), thereby improving the contextual information captured by CBAM and enhancing the model's ability to capture spatial relationships within local features. Additionally, we employ SWA within the bottleneck layer of the Residual U-Net to further enhance the model's performance. To evaluate our approach, we perform experiments using two widely used datasets of breast ultrasound images and the obtained results demonstrate its capability in accurately detecting tumors. Our approach achieves state-of-the-art performance with dice score of 74.23% and 78.58% on BUSI and UDIAT datasets, respectively in segmenting the breast tumor region, showcasing its potential to help with precise tumor detection. By leveraging the power of deep learning and integrating innovative attention mechanisms, our study contributes to the ongoing efforts to improve breast cancer detection and ultimately enhance women's survival rates. The source code of our work can be found here: https://github.com/AyushRoy2001/DAUNet. [ABSTRACT FROM AUTHOR]