3 results about "Image noise reduction" patented technology

A diagnostic imaging system in an example comprises a high frequency electromagnetic energy source, a detector, a data acquisition system (DAS), and a computer. The high frequency electromagnetic energy source emits a beam of high frequency electromagnetic energy toward an object to be imaged. The detector receives high frequency electromagnetic energy emitted by the high frequency electromagnetic energy source. The DAS is operably connected to the detector. The computer is operably connected to the DAS and programmed to employ a threshold to trigger a filter operation on a pixel, in a basis material decomposition (BMD) image of a plurality of BMD images, through comparison of an actual noise ratio between a pair of BMD images, of the plurality of BMD images, to a theoretical BMD noise ratio value. The computer operably connected to the DAS is programmed to employ a correlation in noise distribution of the plurality of BMD images to reduce image noise in the plurality of BMD images. The computer operably connected to the DAS is programmed to realize an adaptive algorithm through employment of an exponential correction function of a difference between the actual noise ratio and the theoretical BMD noise ratio value. The computer operably connected to the DAS is programmed to employ the adaptive algorithm to reduce the image noise in the plurality of BMD images.

The invention relates to an image noise reduction method and a system. The method comprises the steps of: converting an image to be processed into at least one image sub-layer by a client side, and transmitting the image sub-layers to a cloud computing system; performing noise reduction on the image sub-layers after the image sub-layers are received by the cloud computing system, and returning the image sub-layers subjected to noise reduction to the client side by the cloud computing system; reconstructing the image sub-layers subjected to noise reduction into an image after the image sub-layers subjected to noise reduction are received by the client side by the client side, thereby obtaining the image subjected to noise reduction. Since the image is layered, the complex image is decomposed into multiple image sub-layers for noise reduction and noise of each image sub-layer is reduced simultaneously via cloud computing, high-complexity computation can be performed to obtain high-quality images with noise reduced in the case of limited client side resources, and current demands for rapid and efficient image noise reduction can be satisfied in the case of increase of the image resolution and the number of images.