Stereo Image Quality Assessment Based On Binocular Energy

Image quality assessment is one of the hot topic of current research. Due the Internet is highly developed, get knowledge from picture or video has become a habit in our daily life. Image quality evaluation of the existing methods cannot be well applied to stereoscopic images. Recently, there are many scientists join the field of stereoscopic image quality assessment in order to get a better result. Here are main works of this paper:Introduce the background and significance of the stereoscopic image quality assessment that scholars has made. By understanding other people’s algorithm can give us more inspiration with ideas.No-reference stereoscopic image quality assessment based on Gabor filters. Due to Gabor filtering process is more similar with Human Visual System. First, we get the information of amplitude and phase through filtering stereoscopic image by Gabor filters, then evaluation features of the stereoscopic image, at last use support vector regression (SVR) to get the result.No-reference stereoscopic image quality assessment method based on wavelet transform. Firstly, evaluate the "Cyclopean" images are gained by using Gabor and SSIM algorithm based on the left and right image pairs, and then the sub-band energy of the left and right stereoscopic image and the "Cyclopean" image are calculated by wavelet decomposition, at last the relationship model between perceptual features of 3D image and subjective scores are built by SVR. Experimental results show that our proposed new method is better consistent with subjective assessment, and have a better accord with Human Visual System.No-reference stereoscopic image quality assessment method based on Dual tree complex wavelet transform (DT-CWT), with respect to the wavelet transform, DT-CWT get more direction and sub-band coefficient. And then use sub-band coefficients obtained was processed to extract features. Eventually we tested through training, to calculate the image we get an objective score. PLCC and SROCC are two indicators for us to measure the quality of the algorithm. Through continuous experimentation and testing on LIVE3D_I libraries and LIVE3D II libraries. The result shows that the metric be consistent with subjective evaluation.