An effective framework for passive-blind color image tampering detection is presented in this paper. The proposed image statistical features are generated by applying Markovian rake transform to image luminance component. Markovian rake transform is the application of Markov process to difference arrays which are derived from the quantized block discrete cosine transform 2-D arrays with multiple block sizes. The efficacy of thus generated features has been confirmed over a recently established large-scale image dataset designed for tampering detection, with which some relevant issues have been addressed and corresponding adjustment measures have been taken. The initial tests by using thus generated classifiers on some real-life forged images available in the Internet show signs of promise of the proposed features as well as the challenge encountered by the research community of image tampering detection.