Image processing system, image processing method and program for image processing

An image processing method which comprises:

• computing a geometric transformation parameter which represents a geometric transformation of an image from a previous frame to a current frame and calculating a level of reliability in the geometric transformation parameter;
• generating a prediction of a super-resolution image in the current frame by transforming a super-resolution image generated in the previous frame on the basis of the geometric transformation parameter;
• generating the super-resolution image in the current frame by calculating a low-resolution image in the current frame from the prediction result of the super-resolution image in the current frame by a simulation, calculating a difference between a low-resolution image, which is an input image for the current frame, and the low-resolution image calculated by the simulation, and calculating a weighted mean between a result of upsampling the difference to the same number of pixels as that of the super-resolution image and the prediction result of the super-resolution image in the current frame;
• and
• calculating the weighted mean on the basis of the level of reliability in the geometric transformation parameter in calculating the weighted mean.

A non-transitory computer readable information recording medium storing a program which, when executed by a processor, performs a method comprising:

• computing a geometric transformation parameter which represents a geometric transformation of an image from a previous frame to a current frame and calculating a level of reliability in the geometric transformation parameter;
• generating a prediction of a super-resolution image in the current frame by transforming a super-resolution image generated in the previous frame on the basis of the geometric transformation parameter;
• and
• generating the super-resolution image in the current frame by calculating a low-resolution image in the current frame from the prediction result of the super-resolution image in the current frame by a simulation, calculating a difference between a low-resolution image, which is an input image for the current frame, and the low-resolution image calculated by the simulation, and calculating a weighted mean between a result of upsampling the difference to the same number of pixels as that of the super-resolution image and the prediction result of the super-resolution image in the current frame,
• wherein, when generating the super-resolution image, calculating the weighted mean on the basis of the level of reliability in the geometric transformation parameter.

An image processing system comprising:

• a hardware including a processor and a memory,
• the hardware configured to:
• compute a geometric transformation parameter which represents a geometric transformation of an image from a previous frame to a current frame;
• generate a prediction of a super-resolution image in the current frame by transforming a super-resolution image generated in the previous frame on the basis of the geometric transformation parameter;
• and
• generate the super-resolution image in the current frame by:
• calculating a low-resolution image in the current frame from the prediction result of the super-resolution image in the current frame by a simulation,
• calculating a difference between a low-resolution image, which is an input image for the current frame, and the low-resolution image calculated by the simulation, and
• calculating a weighted mean based on the difference.