Table 1 Pseudocode for the SS2D calculation process.

Input: Feature map F

Step1: Divide F into blocks of size pxq

Step2: Perform IE-Scan to obtain sequence X(t), t \(\in\){1}

perform cross-Scan to obtain sequence X(t), t \(\in\){2,3,4,5}

Step3: Process sequence X through the S6 module to obtain the output sequence Y(t)

      \(Y\left( t \right) = {\text{S}}6\left( {X\left( t \right)} \right),{ }t \in { }\left\{ {1,2,3,4,5} \right\}\)

Step4: Reverse the order of the 5 output sequences Y(t) to restore the intermediate features F_t to the same size as F

Step5: Fusion the 5 intermediate features F_t by summing along the channel dimension to obtain the output feature \({F}{\prime}\)

      \(F^{\prime} = {\text{Sum}}\left( {F_{t} } \right),{ }t \in { }\left\{ {1,2,3,4,5} \right\}\)

Output: Feature map \({F}{\prime}\)