The emerging need for automation by using artificial intelligence with machine learning has increased the digital data in the form of images and videos. Images and videos in raw uncompressed format have very large data. Transmission and storage of such large data is not easy. So, compression techniques are adopted. The innovations in multimedia applications and the increasing usage of its products have led to shortage of both bandwidth and memory. Therefore, to save the memory and to make transmission of data more convenient compression is the need which is achieved by identifying redundancy in data. Depending on the energy levels required to reconstruct the image, some DCT coefficients are discarded and PSNR can be used to decide the coefficients sufficient to reconstruct the image. Research work is more concentrated on using quantisation matrix for compression which is standardised by JPEG. In the proposed algorithm after transforming the image to frequency domain the coefficients required to represent the image can be reduced and the quality of the image can be measured by using both subjective and objective criteria. The proposed algorithm is applied for dimensions of 256x256 image and 128x128 image and its CR and PSNR is calculated and tabulated.