Medical Image-Based Cancer Diagnosis and Detection Using Deep Learning

A prominent area of focus in the sector of computervision and artificial intelligence is the use of deep learning technologies to achieve cancer diagnosis using medical images. On top of the inherent complexity and specificity of medical imaging, the fast evolution of deep learning algorithms has made cancer diagnosis an extremely time-sensitive and precise task. As the top cause of death worldwide, cancer detection is a major obstacle for scientists and medical professionals. It is essential to diagnose cancer early, but current approaches frequently involve intrusive procedures and lengthy analysis, therefore there is a need for better, more efficient alternatives. In order to overcome these obstacles, this research employs AI-based methods for automated cancer detection, with an emphasis on deep learning models are among the Convolutional Neural Networks (CNNs) that have been tested on cancer picture datasets. Image segmentation is the first step in contour feature extraction, which involves the calculation of parameters like area, epsilon, and perimeter. After putting the models through their paces, DenseNet121 emerged victorious with a validation accuracy(Acc) of 99.96%, a loss of 0.0015, and RMSE values of 0.026056 for training and 0.035826 for validation-the lowest of the bunch. Among these findings, DenseNet121 stood out as the most effective model, demonstrating the potential of AIbased strategies to enhance cancer diagnosis accuracy.