Joint Dequantization and Contrast Enhancement of JPEG Images with Poor Lighting Using Graphs PROJECT TITLE : Graph-based Joint Dequantization and Contrast Enhancement of Poorly Lit JPEG Images ABSTRACT: The lossy compression of JPEG images results in images with low contrast and coarse quantization artefacts in low-light situations. Compression artefacts would amplify if dequantization and contrast augmentation were performed sequentially, resulting in poor visual quality. A graph Signal Processing (GSP) framework can be used to dequantize and improve these images at the same time, taking advantage of recent advances in GSP. This is done by defining a generalised smoothness prior and a signed graph smoothness prior for each observed pixel patch based on their unique signal characteristics in the context of Retinex theory. Low-pass filtering in the dual graph domain is used to compute robust edge weights for each graph given only a transform-coded picture patch. APG techniques are used to compute the illumination and reflectance components for each patch, using backtracking line search for even more speedup in the transform domain. It is clear from the experimental results that our generated photos outperform the current state of the art significantly in terms of quality assessment. Did you like this research project? To get this research project Guidelines, Training and Code... Click Here facebook twitter google+ linkedin stumble pinterest From a Single Photograph, Graph-Based Blind Image Deblurring Face Sketch Synthesis with a Graph-Regularized Locality-Constrained Joint Dictionary and Residual Learning