When an image is supposed to have been transformed by a process like image enhancement or lossy image compression for storing or transmission, it is often necessary to measure the quality of the distorted image. This can be achieved using an image processing method called “quality criterion”. Such a process must produce objective quality scores in close relationship with subjective quality scores given by human observers during subjective quality assessment tests. In this paper, an image quality criterion is proposed. This criterion, called C4, is fully generic (i.e., not designed for predefined distortion types or for particular images types) and based on a rather elaborate model of the human visual sys- tem (HVS). This model describes the organization and operation of many stages of vision, from the eye to the ventral and dorsal pathways in the visual cortex. The novelty of this quality criterion relies on the extraction, from an image rep- resented in a perceptual space, of visual features that can be compared to those used by the HVS. Then a similarity metric computes the objective quality score of a distorted image by comparing the features extracted from this image to features extracted from its reference image (i.e., not distorted). Results show a high correla- tion between produced objective quality scores and subjective ones, even for images that have been distorted through several different distortion processes. To illus- trate these performances, they have been computed using three different databases that employed different contents, distortions type, displays, viewing conditions and subjective protocols. The features extracted from the reference image constitute a reduced reference which, in a transmission context with data compression, can be computed at the sender side and transmitted in addition to the compressed image data so that the quality of the decompressed image can be objectively assessed at the receiver side. More, the size of the reduced reference is flexible. This work has been integrated into freely available applications in order to formulate a practical alternative to the PSNR criterion which is still too often used despite its low cor- relation with human judgments. These applications also enable quality assessment for image transmission purposes.