Improvising Low Contrast Malaria Images Using Contrast Enhancement Techniques on Various Color Models
Doni Setyawan, Retantyo Wardoyo, Moh Edi Wibowo, E. Elsa Herdiana Murhandarwati · 2022 Seventh International Conference on Informatics and Computing (ICIC) · 2022
The digital malaria image acquisition may result in low-contrast images. The low contrast images make it difficult to visualize and analyze the morphological features among plasmodium species, which can increase the false diagnosis rate. Therefore, various types of contrast improvement methods have been proposed to increase the contrast of malaria images. Comparison of these methods is essential to determine the appropriate technique for improvement and further processing. This study applied and analyzed the HE, AHE, CLAHE, and GE methods in the grayscale and green channel malaria image and GCS, LCS, MGCS, and MLCS methods in color malaria images. In grayscale malaria images, based on MSE and PSNR measurements, the best results were obtained using the CLAHE, GE, AHE, and HE methods, respectively. In the green channel malaria image, based on the MSE measurement, the smallest error was produced by the CLAHE method, while based on the PSNR measurement, the best quality was obtained by the GE method. CLAHE and GE can visually clarify the texture and morphology of plasmodium and erythrocytes with minimal noise. In color images, based on PSNR and AMBE measurements, the GCS and MGCS methods provide better color quality and preservation than the LCS and MLCS methods. Malaria images with good contrast are expected to help visually examine malaria and facilitate the segmentation process in an automatic malaria diagnosis system.