Scene change detection using sum of variance and estimated picture encoding cost

A video processing device comprising:

• an encoder to one of encode or transcode a video stream;
• a complexity estimation module to determine a first sum of variances metric and a first estimated picture encoding cost metric for a first picture of the video stream;
• a scene analysis module to determine a first threshold based on a first statistical feature for sum of variance metrics of a set of one or more pictures preceding the first picture in the video stream and based on a second statistical feature for the sum of variance metrics of the set of one or more pictures and determine a second threshold based on a first statistical feature for estimated picture encoding cost metrics of the set of one or more pictures and based on a second statistical feature for the estimated picture encoding cost metrics of the set of one or more pictures;
• wherein the encoder modifies at least one of a bit rate or a quantization parameter of the one of the encoding or transcoding of the video stream responsive to the scene analysis module identifying a scene change as occurring at the first picture based on the first sum of variances metric, the first estimated picture encoding cost metric, the first threshold, and the second threshold;
• wherein the scene analysis module refrains from including a sum of variances metric and an estimated picture encoding cost metric of a picture of the set in determining the first statistical features responsive to determining the picture of the set is a repeat picture, wherein the picture is determined to be a repeat picture responsive to determining that the sum of variances metric of the picture of the set is more than a threshold multiple of the estimated picture encoding cost metric of the picture of the set;
• wherein the first statistical feature for the sum of variance metrics of the set of one or more pictures comprises a mean sum of variances metric;
• wherein the second statistical feature for the sum of variance metrics of the set of one or more pictures comprises a mean squared error for the sum of variance metrics of the set of one or more pictures;
• wherein the first statistical feature for the estimated picture encoding cost metrics of the set of one or more pictures comprises a mean estimated picture encoding cost metric;
• and
• wherein the second statistical feature for the estimated pictured encoding costs of the set of one or more pictures comprises a mean squared error for the estimated picture encoding cost metrics of the set of one or more pictures.

A non-transitory computer readable medium storing a set of executable instructions, the set of executable instructions comprising instructions to manipulate at least one processor to:

• perform one of an encoding or a transcoding of a video stream;
• determine a first sum of variances metric and a first estimated picture encoding cost metric for a first picture of the video stream;
• determine a first threshold based on a first statistical feature for sum of variance metrics of a set of one or more pictures preceding the first picture in the video stream and based on a second statistical feature for the sum of variance metrics of the set of one or more pictures and determine a second threshold based on a first statistical feature for estimated picture encoding cost metrics of the set of one or more pictures and based on a second statistical feature for the estimated picture encoding cost metrics of the set of one or more pictures;
• refrain from including a sum of variances metric and an estimated picture encoding cost metric of a picture of the set in determining the first statistical features responsive to determining the picture of the set is a repeat picture, wherein the picture is determined to be a repeat picture responsive to determining that the sum of variances metric of the picture of the set is more than a threshold multiple of the estimated picture encoding cost metric of the picture of the set;
• modify at least one of a bit rate or a quantization parameter of the one of the encoding or transcoding of the video stream responsive to identifying a scene change as occurring at the first picture based on the first sum of variances metric, the first estimated picture encoding cost metric, the first threshold, and the second threshold;
• wherein the first statistical feature for the sum of variance metrics of the set of one or more pictures comprises a mean sum of variances metric;
• wherein the second statistical feature for the sum of variance metrics of the set of one or more pictures comprises a mean squared error for the sum of variance metrics of the set of one or more pictures;
• wherein the first statistical feature for the estimated picture encoding cost metrics of the set of one or more pictures comprises a mean estimated picture encoding cost metric;
• and
• wherein the second statistical feature for the estimated pictured encoding costs of the set of one or more pictures comprises a mean squared error for the estimated picture encoding cost metrics of the set of one or more pictures.