Olution operation to get the Query, Key, and Worth branches. Just after entering the Q

Olution operation to get the Query, Key, and Worth branches. Just after entering the Q branch, the feature map with aa size of C HW was flatbranches. Just after getting into the Q branch, the function map with size of C H W was flattened intotwo-dimensional vector using a size of of N, where N =N = eature map tened into a a two-dimensional vector with a size C C N, where H W. W. Function map Q was Combretastatin A-1 Purity & Documentation transposed to receive a function vector Q’ using a size of N C. Just after the feature Q was transposed to acquire a function vector Q’ with a size of N C. Right after the feature map map entered branch K, the function map using a size of C H W was obtained by way of entered branch K, the function map with a size of C H W was obtained via spatial spatial pyramid pooling to achieve a reduction in dimensionality. The spatial pyramidRemote Sens. 2021, 13, 4532 Remote Sens. 2021, 13, x FOR PEER REVIEW6 of 20 6 ofpooling operationto obtain a reduction in dimensionality. The spatial pyramid module pyramid pooling is shown in Figure five below. The spatial pyramid pooling pooling performedis shown in Figure five under. The spatial pyramid with a window size of n nthe operation the maximum pooling of your input function map pooling module performed to receive the feature map the input feature map n. MAC-VC-PABC-ST7612AA1 custom synthesis Theafeature map with n size of C n n maximum pooling of using a size of C n with window size of a n to obtain the was utilized to represent the sampling benefits of representative anchor points in each location of feature map having a size of C n n. The feature map having a size of C n n was made use of for the origin feature map. Then, each of the feature maps just after the spatial pyramid pooling were represent the sampling results of representative anchor points in each region with the origin flattened and concatenated to receive a feature vector with a size of C S, where S was function map. Then, all of the feature maps following the spatial pyramid pooling have been flattened determined by the size and number of the selected pooling windows. As an example, in this and concatenated to receive a function vector using a size of C S, exactly where S was determined short article, the pooling widow is 1 1, three three, six 6, and eight eight, and S is equal to: by the size and number of the chosen pooling windows. One example is, within this post, the pooling widow is 1 1, three 3, 6 six, = eight 8, and =is equal to: S and n2 S=n1,3,6,8 , , , =Figure 5. Structure of spatial pyramid pooling. Figure 5. Structure of spatial pyramid pooling.Just after the function map, X entered the Query and Crucial branches, along with the feature vectors Right after the function map, X entered the Query and Key branches, as well as the feature vectors Q’ having a size of N C and K’ using a size of C S are matrix multiplied to obtain function Q’ using a size of N C and K’ using a size of C S are matrix multiplied to obtain feature map QK’. Function map QK’ was normalized by SoftMax to obtain the interest map QK. map QK’. Function map QK’ was normalized by SoftMax to acquire the focus map QK. The goal of this was to calculate the relationship among every pixel in function vector The purpose of this was to calculate the connection in between each and every pixel in feature vector Q’ and every pixel in K’. Within this way, we are able to get a feature map of C S size, which Q’ and each pixel in K’. In this way, we can obtain a function map of C S size, which represents the consideration connection among the Query pixel and the function anchor point represents the attention relationship in between the Query pixel and the function anchor point in the Crucial, and repres.