Improving the Texture and Quality of Dermoscopic Melanoma Skin Cancer Images Using Pre Filter Techniques in IRT Images

—The outcomes of the detection mechanism for melanoma skin cancer are very sensitive to the quality of dermoscopic images. Such images are often blurred, noisy, unevenly illuminated, or having rather distorted textures. We present a novel method for texture and quality enhancement of dermoscopic melanoma images with pre-filter techniques specific to Infrared Thermographic (IRT) imaging. Application of such pre-filter algorithms such as Gaussian smoothing, median filtering, and bilateral filtering results in improved clarity, contrast, and texture consistency of the images. The proposed framework was validated on publicly available melanoma datasets, yielding a significant improvement in feature extraction accuracy and melanoma detection performance. This research underscores the importance of preprocessing techniques in achieving reliable diagnostic outcomes in dermoscopy and paves the way for integrating pre-filter methods into clinical melanoma screening workflows.