Abstract: The present disclosure relates to a method for image patch selection for training a neural network for image quality assessment. The method includes receiving an input image and extracting one or more image patches from the input image. The moment of the extracted image patches is measured. There is a decision to accept or decline the extracted image patches according to the measured moment. Additional image patches are extracted until a minimum number, Nmin, of extracted image patches are accepted. Alternatively, selection criteria are adjusted until the minimum number of extracted image patches are accepted. The selected image patches are input into a neural network with a corresponding image quality value of the input image, and the neural network is trained with the image patches and image quality value. Also provided is a method for image quality assessment using a neural network trained as set forth above.

Abstract: A method for distributed video transcoding consists of segmentation process 101, transcoding process 102 and combining process 103. The said segmentation process consists of time division based segmentation 201, spatial division based segmentation 301 and hybrid segmentation 401. The source media is segmented into a number of media segments. These segments are distributed to different processing units 501. Each processing units consists of source reception unit 502, segmentation unit 503, transcoding unit 504, combining unit 505 and finally result storage unit 506 to achieve parallel conversion throughout the process and combined to produce a single transcoded result. The present invention relates to converting an audio-video source media from one format to another within a short duration of time.

Abstract: A method based on CTU level rate-distortion optimization for rate control in video coding which can effectively improve the perceptual rate-distortion performance and coding efficiency is provided. Firstly, a perceptual rate-distortion model is established using a divisive normalization framework, which characterizes the relationship between local visual quality and coding bits. Subsequently, the established perceptual rate-distortion model is applied to overall distortion optimization which is transformed into a global optimization problem and solved with convex optimization algorithms to obtain optimal CTU level coding bit allocation.

Abstract: A machine learning based initial quantization parameter (QP) prediction method, which can effectively optimize RC performance A machine learning framework for initial QP prediction is proposed, where learning labels are built with the criterion of maximizing rate-distortion (RC) performance, which is proved to be much more effective than the QP determination method with the only consideration on sum of the absolute transformed difference (SATD) complexity. Instead of target bits per pixel for intra frame, target bits per pixel for remaining frames is used as sample data to avoid empirically setting intra frame bit allocation, thus improve the prediction accuracy as the real-time updated remaining bits can better reflect the real-time requirements on the level of QPs. In addition, a clipping and decision approach based on the previous initial QP and the target bits per pixel for all remaining frames is proposed, which can help fast QP adaption and quality smoothness.

Abstract: A method based on CTU level rate-distortion optimization for rate control in video coding which can effectively improve the perceptual rate-distortion performance and coding efficiency is provided. Firstly, a perceptual rate-distortion model is established using a divisive normalization framework, which characterizes the relationship between local visual quality and coding bits. Subsequently, the established perceptual rate-distortion model is applied to overall distortion optimization which is transformed into a global optimization problem and solved with convex optimization algorithms to obtain optimal CTU level coding bit allocation.

Abstract: A machine learning based initial quantization parameter (QP) prediction method, which can effectively optimize RC performance. A machine learning framework for initial QP prediction is proposed, where learning labels are built with the criterion of maximizing rate-distortion (RC) performance, which is proved to be much more effective than the QP determination method with the only consideration on sum of the absolute transformed difference (SATD) complexity. Instead of target bits per pixel for intra frame, target bits per pixel for remaining frames is used as sample data to avoid empirically setting intra frame bit allocation, thus improve the prediction accuracy as the real-time updated remaining bits can better reflect the real-time requirements on the level of QPs. In addition, a clipping and decision approach based on the previous initial QP and the target bits per pixel for all remaining frames is proposed, which can help fast QP adaption and quality smoothness.

Abstract: The present disclosure relates to a method for image patch selection for training a neural network for image quality assessment. The method includes receiving an input image and extracting one or more image patches from the input image. The moment of the extracted image patches is measured. There is a decision to accept or decline the extracted image patches according to the measured moment. Additional image patches are extracted until a minimum number, Nmin, of extracted image patches are accepted. Alternatively, selection criteria are adjusted until the minimum number of extracted image patches are accepted. The selected image patches are input into a neural network with a corresponding image quality value of the input image, and the neural network is trained with the image patches and image quality value. Also provided is a method for image quality assessment using a neural network trained as set forth above.

Abstract: An intelligent method and system for expanding and enhancing the color content of video or image to a wider color gamut to make the displayed video or image more pleasing while certain important and sensitive colors such as skin colors are preserved to avoid the perception of visually unnatural image or video artifacts. With the implementation of such method and system, the merits of the wide color gamut display can be fully utilized to provide videos and images of high visual quality.

Abstract: A method and apparatus for converting a standard dynamic range (SDR) video to a high dynamic range (HDR) video. The conversion is adaptive and takes both spatial and temporal information of a current frame and previous frames into consideration such that the majority of pixels falls into the most sensitive regions of human eyes in the target dynamic range, while at the same time, the continuity of luminance is maintained in temporal domain to prevent flickering.

Abstract: An intelligent method and system for expanding and enhancing the color content of video or image to a wider color gamut to make the displayed video or image more pleasing while certain important and sensitive colors such as skin colors are preserved to avoid the perception of visually unnatural image or video artifacts. With the implementation of such method and system, the merits of the wide color gamut display can be fully utilized to provide videos and images of high visual quality.

Abstract: A method and apparatus for converting a standard dynamic range (SDR) video to a high dynamic range (HDR) video. The conversion is adaptive and takes both spatial and temporal information of a current frame and previous frames into consideration such that the majority of pixels falls into the most sensitive regions of human eyes in the target dynamic range, while at the same time, the continuity of luminance is maintained in temporal domain to prevent flickering.

Abstract: A method for distributed video transcoding consists of segmentation process 101, transcoding process 102 and combining process 103. The said segmentation process consists of time division based segmentation 201, spatial division based segmentation 301 and hybrid segmentation 401. The source media is segmented into a number of media segments. These segments are distributed to different processing units 501. Each processing units consists of source reception unit 502, segmentation unit 503, transcoding unit 504, combining unit 505 and finally result storage unit 506 to achieve parallel conversion throughout the process and combined to produce a single transcoded result. The present invention relates to converting an audio-video source media from one format to another within a short duration of time.