Breast cancer ranks first among the types of cancer affecting women worldwide. However, early detection can lead to effective treatment. Developing a system to make the decision on the benign or malignant nature of the tumour will help radiologists to establish a precise diagnosis in order to manage patients presenting the pathologies. The objective of this study is to develop a method that makes it possible to recognize the malignant or benign nature of breast cancer in a mammographic image, using the neural network. The method we developed is based on the neural network. Indeed, we extracted the characteristics of the mammographic images having undergone pre-processing and the detection of regions of interest by the multiscale product. These characteristics are extracted, in one hand by the Gray Level Co-occurrence Matrix (GLCM) and in the other hand the Gray Level Run Length matrix (GLRLM). The extracted characteristics constitute the data at the input of a neural network (the pattern net). The mammographic images from the MIAS database were used as a learning basis and recognition basis. This development allowed us to classify objects in a region of interest as malignant or benign. The results of the proposed method showed sensitivity, specificity and an area under the curve all equal to 1 for images labelled malignant and therefore cancerous. For images labelled benign, the sensitivity is equal to 0.8, the specificity is equal to 1, and the area under the curve is equal to 0.88. Thus, the results highlighted the effectiveness of the method we proposed. Compared to the results of the literature (recent state of the art), we can say that the method we proposed is the most efficient in terms of evaluation criteria. The most contributions of this work are the successful using of the characteristics of mammographic images (characteristics issues from the segmentation and multiscale product of 2-Dimensional continuous wavelet coefficient) extracted by GLCM and GLRLM and used as the input of a neural network to recognize and classify as benign or malignant the mammographic images.
