An Analytical Approach for Reconstruction of Cosmetic Surgery Images using EUCLBP and SIFT
S K Shiji, Krishnaveni S H · International Journal of Electrical and Electronics Engineering · 2022
Plastic surgery is a surgical procedure with outcomes often opposite to facial ageing. Face recognition algorithms find it difficult to anticipate such non-uniform face transformations because any changes brought on by plastic surgery procedures happen quickly. In this paper, the performance of face reconstruction is compared with that of two commonly used feature extractors, Extended Uniform Circular Local Binary Pattern (EUCLBP) and Scale Invariant Feature Transform (SIFT). An open plastic surgery facial dataset of 1800 before and after operation picture samples for 900 human face photographs has been used to test all algorithms. For each human subject, two front-facing image samples with appropriate luminance and neutral motions are gathered; the first is taken before, and the second is obtained after surgery. The proposed method is validated using parameters like accuracy, sensitivity, specificity, F-Score, G-mean and Precision. These results consistently show superior performance and high identification accuracy of 97% in combination with using both feature extractors EUCLBP+SIFT rather than using any one [EUCLBP or SIFT]alone.