36results about How to "Increase information entropy" patented technology

The invention discloses a medical digital subtraction image fusion method based on ridgelet transformation, which comprises the following steps: (1) respectively performing ridgelet transformation on two images to obtain a ridgelet transformation matrix; (2) performing fusion processing; ( 3) Carry out ridgelet inverse transform to the ridgelet transformation matrix after fusion, and the reconstructed image obtained is the image after fusion; it is characterized in that: in step (1), the initial judgment threshold and the step size are set, and the process is carried out accordingly. Ridgelet transform, use the inverse transformation to reconstruct the fusion image, calculate the information entropy of the fusion image, use the dynamic fuzzy method to change the judgment threshold according to the step size and repeat the above operation, use the judgment threshold corresponding to the maximum information entropy as the final judgment threshold, and obtain the step (1 ) described ridgelet transformation matrix. The invention can effectively improve the information entropy of the fusion image, reduce the root mean square error rate, and the performance is better than other traditional fusion methods. The time complexity of the algorithm is low, and the results obtained are relatively accurate, which greatly enriches the texture details of medical images.

The invention discloses a method for dynamically calibrating image light intensity on the basis of environmental factors. The method includes the steps of reversing an image to distinctly display key areas; segmenting the image into NxN sub-blocks for processing; processing the sub-blocks via power transformation; processing the sub-blocks via grayscale mapping transformation; subjecting the sub-blocks to histogram processing; processing the sub-blocks via dynamic fuzzy histogram equalization; respectively processing the sub-blocks via masking and filtering by a hybrid space processing method; iterating every sub-blocks of the NxN sub-blocks by a Lucy_Richardson filter; respectively calculating to optimize the power transformation step and the dynamic fuzzy histogram equalization step of the sub-blocks by a fuzzy logic theory; and reversing the image again and outputting the final image, wherein the N is an integer nine, the maximum information entropy is judged as the optimum optimizing standard. By the aid of the method, information entropy of the image is effectively increased, and good light calibrating effect is obtained.

The invention discloses a multi-channel data processing method, device, electronic equipment and storage medium. Including: sending the inconsistent data independently generated by the corresponding functional modules of each channel in the combined failsafe architecture to each other channel in the combined failsafe architecture; sending the inconsistent data independently generated by the corresponding functional modules of each channel and receiving Inconsistent data independently generated by the corresponding functional modules of each other channel in the combined fail-safe architecture, according to the preset data processing method, each channel in the combined fail-safe architecture is processed independently, and each channel is independent The processed new data; the subsequent input data of each channel is determined according to the new data independently processed by each channel. It can improve the identification ability and applicability of the combined fault safety technology to the abnormality of the channel, play the role of the combined fault safety technology, improve the safety, and ensure the independence of the channel.