Cervical cancer is the second most common type of cancer in women from underdeveloped countries. For the analysis of cervical lesions, the colposcopy with biomarkers, such as acetic acid and lugol are used to visually differentiate normal tissues and pathologies of the cervix. This procedure makes use of different electronic devices such as cameras, screens and lamps to observe the vaginal cavity. The lamps reflecting on the cervix generate highly bright areas, which make it difficult to diagnose them as normal or abnormal tissues. Objective: The purpose of this work is to present a model of reconstruction of images of the cervix that are affected with brightness. We show a resultant image reconstructed without brightness. Methodology: We propose an algorithm for the reconstruction of cervical images with biomarkers. For which we build a prototype of cervix, with expanded polystyrene, plasticine and enamel. We established three images captured from a constant position, varying the location of the light focus to achieve the brightness displacement over the surface, artificial vision algorithms were applied for the recognition of pixels with brightness. Then, we choose a base image on which the reconstruction process takes place. Results: We got an image with the most complete information of the zones with their respective colors and other characteristics. A recovery percentage equal to or greater than 99% was achieved, proportional to the number of catches taken in the first step. This facilitated the diagnosis to gynecologic specialists to the Cints project.