Optimized Breast Cancer Diagnosis through Medical Image Processing: Integration of Multiwavelet Transform, Gamma Correction, and Hybrid Deep Convolutional Neural Networks

Breast cancer occurs when cells or tissues of the breast grow abnormally.On a global scale, Globally, cancer is a common and deadly type of disease that most affects women.These studies provide a new paradigm for skillful deep recognition of preprocessed images using deep neural networks.At this stage, the paper presents an efficient BreastNet model for breast cancer detection containing 48 layers, applied to three datasets namely breast ultrasound images (BUST), which includes 780 images classified into benign, malignant, normal and invasive ductal carcinoma (IDC) and BreakHis_v1_400X.Breast cancer images are processed using different feature extraction vectors.Medical image processing is a crucial step in adjusting the brightness and contrast of images.Gamma correction can help preserve detail in the darkest and brightest areas of an image.In addition to its combination of complex-frequency features, Geronimo-Hardin-Massopust (GHM) multiwave conversion is a type of multiwave conversion that offers many advantages for image processing tasks.By capturing low-frequency (soft areas) and high-frequency (edges and textures) components of an image more accurately, Discrete Wavelet Transform (DWT) processing can also In order to represent an image more efficiently, it is necessary to reduce the amount of data required while retaining important visual details.The proposed model, BreastNet, improved the classification accuracy performed on the proposed dataset by 95.88%, 93.30%, and 91.16%.The proposed model has a better execution rate than AlexNet, VGG16, Googlenet, Resnet18, and Resnet50 models.