Nowadays, the digital image processing technology is applied for many fields. In this thesis, a new method is introduced to enhance handwritten document images such as historical document that reflects the creative achievements in different historical periods, records historical events, the great devotion of the heroes, scientists and the famous cultural, etc.
We proposed a new method for handwriting documents denoising, called Adaptive Directional Lifting Wavelet Transform (ADLWT), which differs from others in the two aspects: Firstly, using contrast limited adaptive histogram equalization to equalize the contrast of an image by cutting histogram at some threshold, after that equalization is applied. The Contrast Limited Adaptive Histogram Equalization (CLAHE) is different from other methods, i.e., it can improve the contrast of an image by applying a method, in which not only the contrast of an image is improved by applying for small data areas, but also the enhancement function can be applied through all neighborhood pixels. The contrast in the homogeneous region can be limited hence the noise amplification can be avoided. These produce the distribution of used grey values and hence make hidden features of the image more visible. Secondly, the improvement for the image is implemented by using adaptive directional lifting-based discrete wavelet transform enhancing operation for the foreground and interfering strokes, respectively. As a result, this method not only removes the interfering strokes or visible watermarks in background information but also significantly increases the readability of handwritten document images.
