Breast cancer is one of the most prevalent cancers. Before initiating treatment, the phase of breast histopathology images' segmentation is crucial for obtaining an accurate diagnosis. The effectiveness of segmentation is frequently dependent on enormous training datasets accompanied by high-quality human annotations. However, the annotation process is labor-intensive, costly, and consumes much time. This paper proposes a novel color-detection-based method for automatically annotating breast cancer histopathology images. We also build a semantic segmentation model for breast cancer histopathology images based on deep learning using the UNet architecture allowing the pathologist to make immediate and accurate diagnoses.