Bear in mind that none of these objective measures are particularly good at predicting human visual response to image quality. Sometimes PSNRs vary wildly between two almost indistinguishable images; similarly you can have two images with the same PSNR where there is a very obvious difference in quality. The structural similarity index measurement (SSIM) and some of its variations are generally considered better from this perspective, but still not perfect models for human perception.
PSNR high means good quality and low means bad quality. PSNR is using a term mean square error (MSE) in the denominator. So, low the error, high will be the PSNR.
Peak Signal to Noise Ratio ( PSNR) and Mean Square Error (MSE) are used to comparing the squared error between the original image and the reconstructed image.There is an inverse relationship between PSNR and MSE. So ahigher PSNR value indicates the higher quality of the image (better).
Bear in mind that none of these objective measures are particularly good at predicting human visual response to image quality. Sometimes PSNRs vary wildly between two almost indistinguishable images; similarly you can have two images with the same PSNR where there is a very obvious difference in quality. The structural similarity index measurement (SSIM) and some of its variations are generally considered better from this perspective, but still not perfect models for human perception.
thanks for your valuable answer's....finally i find it simple concept of PSNR..if PSNR is high better for Compression and Stegnography..but encryption concept PSNR very low is better.
UQI,SSIM, are better than psnr for assessing the image quality , a simple scale in the image pixels can cause great reduction in PSNR. PSNR as such is not reliable for quality assessment
PSNR- Peak signal to noise ratio. Calculated usually in logarithmic (dB) scale is a metric use to measure the quality of any image reconstructed, restored or corrupted image with respect to its reference or ground truth image. It is a full reference image quality measure defined as the maximum value of maximum signal power with respect to MSE (Mean square error) assumed as noise power. For 8-bit image maximum signal power is (2^(8-1) )^2 i.e. 255^2. Similarly MSE can be calculated as the square difference between reference image and reconstructed/restored image. Thus a higher value of PSNR indicates that the image is of higher quality and vice-versa. A 20 dB or higher PSNR indicates that the image is of good quality.
PSNR is an image quality estimator after compression or some modification to the image, just like SSIM, MSE and etc. Thus its technically incorrect to say if PSNR high => image quality is high (although it is true for extremely high PSNR).
Instead of thinking with respect to the quality level, it is more helpful to think in terms of what it is measuring: PSNR is related to sum of squared function, i.e., PSNR value is higher for evenly distributed modification than sparsely distributed modification.
Thus instead of modifying one pixel value by a large amount, modifying many pixels by very small amount will result in higher PSNR. (This is based on one of the human visual system model)
PSNR high means: Mean square error between the original image and reconstructed image is very low. It implies that the the has been properly restored. In the other way, the restored image quality is better.
If PSNR value is low; the quality of the restored image is very bad.
MSE is zero means no noise is present in the signal .There peak noise to signal ratio have no importance.
One problem with mean-squared error is that it depends strongly on the image intensity scaling. A mean-squared error of 100.0 for an 8-bit image (with pixel values in the range 0-255) looks dreadful; but a MSE of 100.0 for a 10-bit image (pixel values in [0,1023]) is barely noticeable.
If the PSNR value is high, the picture quality is better, 40 is best, picture quality is slightly lower at 30 db, noise is increasing at 20 dB. 10 dB noise increase continues, 0 dB picture image is completely noise.
PSNR acronym for Peak Signal to Noise Ratio which represents
a metric widely used in signal processing to measure the quality of a signal by calculating the ratio between the original signal and the noise. Expressed in decibels, the bigger it is (> 35 dB), the better the quality.
PSNR is measured regarding another image. That is, one is original and another is modified image. If the PSNR is high, it indicates that your modified image is very close to original one. However, PSNR works for intensity comparison. It does not provide any structural information. Hence, you can also apply SSIM or UIQI methods to compare image quality.
Although the question is made 5 years ago, the answer may be still interesting.
PSNR is not a good model for human perception. But as long as the content is the same, it is a very good measure to assess and compare the quality variations. I would recommend you to read this paper:
"Scope of validity of PSNR in image/video quality assessment", Q. Huynh-Thu, M. Ghanbari
PSNR only quantifies the quality of a reconstructed or corrupt image with reference to the ground-truth. It not compared to the HVS. So a better evaluation metric will be SSIM which evaluates the structure of the image.The SSIM of a reconstructed image to ground-truth image is always 1, and with a value close to 1, one can tell the image is of good quality.