Researchers have developed a systematic method of using image-processing techniques to assess the complexity of the background of overhead guide and street name signs under nighttime driving conditions. These techniques are used to extract image properties such as entropy, contrast, energy, homogeneity, the number of saturation pixels, the edge ratio, and the number of objects, all of which are considered potential factors for evaluating background complexity. The researchers combined these factors with ratings of images by human survey participants to develop a multiple linear regression model that could be used by practitioners to evaluate the background complexity of overhead guide and street name signs under nighttime conditions. Because of the small number of samples in the data sets, bootstrapping, a resampling method, was employed to improve the performance of the proposed model. High consistency between the results of the proposed model and the empirical results from the survey demonstrated that the model performed well in analyzing the complexity of the background of traffic signs. Practitioners can use this model to identify overhead guide and street name signs that have highly complex backgrounds and may require sign lighting, supplemental signs, or relocation to minimize driver difficulty in detecting and obtaining information from the signs.