In recent years, many studies have made progress in identifying bias in facial recognition systems. However, existing works recognize that bias in facial recognition still exists, leading to false identifications, arrests, and agony for our citizens. Current systems cannot effectively detect African American men and women between the ages of 18 and 30. This group faces bias in different facial recognition system algorithms, likely due to certain underlying factors. Factors include the image quality of datasets and untrained or insufficiently trained algorithms to accurately detect faces in images. These factors have led investigators to introduce a new principle in understanding how race and demographic bias can be found in facial recognition algorithms. Through our research, we propose a new method that can be implemented in facial recognition algorithms during the detection phase. This paper introduces this new method to decrease unfairness during the classification phase of datasets. The fairness measure is calculated during identification of facial landmarks used to detect parts of the face with coordinates. By applying a RYB (red, yellow, blue) pallet of colors when detecting areas of the face instead of the more conventional RGB (red, green, blue) pallet. Our research indicates RYB colors may improve the equality of images, which can decrease bias and improve detection fairness in algorithms for captured images of African Americans ages 18–30.