Identification of even minor modifications in images belongs to an intensifying branch of multimedia forensics. Numerous solutions have been identified and employed to detect the common attacks made to tamper the image. Copy-Move is most commonly used attack chosen by forger to tamper the image. Detection of copy-move is mainly carried out by using block matching and key-point based methods. This paper proposes hybrid framework for copy-move tamper detection. This hybrid approach combines traditional Discrete Cosine Transform and Principal Component Analysis (DCT-PCA) based feature extraction. The proposed method is evaluated for 8x8 and 16x16 block size and effectively works for both block sizes. K-means clustering is used for block division which is faster than other techniques block clustering algorithms. Results show that, overall accuracy of the system is increased by applying K-means clustering approach. Experimental analysis indicates that system can handle various attacks successfully such as Translation, Rotation, Distortion and Combination of two or more transformations. Proposed system gives 90% success rate with minimum execution time.