In this paper, a color image reduction algorithm is implemented that makes use of averaging functions to reduce the images without decomposing them to RGB channels. Penalty functions are used to find out the most optimal averaging function that provides least error between original and reconstructed image. Simulation results show that the implemented algorithm reduces the size of the color images so that can be transmitted at low cost. The average values of MSE, MAE and SSIM in case of reduction of the 12 test images used our simulations are 250.73, 8.47 and 0.7217 respectively. Moreover, this scheme also does well when the test image is corrupted with noise. Technical Journal, University of Engineering and