To search, Click below search items.


All Published Papers Search Service


Eigentransformation with Error Regression Model for Face Hallucination


Kanjana Boonim, Parinya Sanguansat


Vol. 11  No. 5  pp. 76-83


Generally, a high-resolution (HR) face image is reconstructed only from low-resolution (LR) face image. However, previous researches neglected to gain benefits from error of face reconstruction. This paper proposes a new face hallucination technique for face image reconstruction using Eigentransformation with error regression model. In order to improve the performance of facial image reconstruction, the error information is included in our framework to correct the final result. In this way, the regression analysis is used to find the error estimation which can be obtained from the existing LR in eigen space. Our framework can work with both grayscale and color images. However, to handle with color images, each color channel in RGB model must be separately processed. The framework consists of learning and hallucinating process. In learning process is from the mistakes in reconstruct face images of the training dataset by Eigentransformation, then finding the relationship between input and error by regression analysis. In hallucinating process uses normal method by Eigentransformation, after that the result is corrected with the error estimation. Experimental results from the well-known facial databases show that the final resolution and quality are greatly enhanced over the sole Eigentransformation method.


Eigentransformation, error regression model, Face hallucination, Principal Component Analysis (PCA)