Skip to main content

Research Repository

Advanced Search

Image Enhancement for Scanned Historical Documents in the Presence of Multiple Degradations

Suleiman, Farouk

Image Enhancement for Scanned Historical Documents in the Presence of Multiple Degradations Thumbnail


Farouk Suleiman



Historical documents are treasured sources of information but typically suffer from problems with quality and degradation. Scanned images of historical documents suffer from difficulties due to paper quality and poor image capture, producing images with low contrast, smeared ink, bleed-through and uneven illumination. This PhD thesis proposes a novel adaptative histogram matching method to remove these artefacts from scanned images of historical documents. The adaptive histogram matching is modelled to create an ideal histogram by dividing the histogram using its Otsu level and applying Gaussian distributions to each segment with iterative output refinement applied to individual images. The pre-processing techniques of contrast stretching, wiener filtering, and bilateral filtering are used before the proposed adaptive histogram matching approach to maximise the dynamic range and reduce noise. The goal is to better represent document images and improve readability and the source images for Optical Character Recognition (OCR). Unlike other enhancement methods designed for single artefacts, the proposed method enhances multiple (low-contrast, smeared-ink, bleed-through and uneven illumination). In addition to developing an algorithm for historical document enhancement, the research also contributes a new dataset of scanned historical newspapers (an annotated subset of the Europeana Newspaper - ENP – dataset) where the enhancement technique is tested, which can also be used for further research. Experimental results show that the proposed method significantly reduces background noise and improves image quality on multiple artefacts compared to other enhancement methods. Several performance criteria are utilised to evaluate the proposed method’s efficiency. These include Signal to Noise Ratio (SNR), Mean opinion score (MOS), and visual document image quality assessment (VDIQA) metric called Visual Document Image Quality Assessment Metric (VDQAM). Additional assessment criteria to measure post-processing binarization quality are also discussed with enhanced results based on the Peak signal-to-noise ratio (PSNR), negative rate metric (NRM) and F-measure.
Image Enhancement, Historical Documents, OCR, Digitisation, Adaptive histogram matching


Suleiman, F. (2024). Image Enhancement for Scanned Historical Documents in the Presence of Multiple Degradations. (Thesis). University of Salford

Thesis Type Thesis
Deposit Date Mar 10, 2024
Publicly Available Date Apr 27, 2024
Award Date Mar 26, 2024


You might also like

Downloadable Citations