Many valuable documents stored on the libraries and to keep the historical document content safe, a binarization approach on document images must be applied. In this study, a modification of Feng algorithm using Max-Mean was proposed. Feng method produced a blurred image and introduced a noise in the resulting image. The proposed method successfully to overcome the Feng method problem by suggesting a maximum threshold value. Based on the result, Max-Mean method able to improve the image quality and reduced the noise. Based on IQA, the proposed method achieved F-measure (74.901), PSNR (16.104), NRM (0.153) and MPM (1.371). Accuracy also shows the increment compare to the Feng result.
