Self-Similarity Superresolution for Resource-Constrained Image Sensor Node in Wireless Sensor Networks

Algorithm 1

Self-similarity regulation scheme.

Input: LR image 𝑌, LR image patches’ size 𝑚 and HR image patches’ size 𝑛, the degradation matrix ℋ.

Output: HR image 𝑋

Step  1.  Extract patches 𝑦_LR∈𝑅^𝑚 from LR image 𝑌, follow the raster-scan order, and start from the upper-left corner

   (some pixel overlap in each direction is allowed).

Step  2.  Recover HR image patches 𝑥_HR iteratively by Steps  2.1 and  2.2, until the maximum iteration times

  or convergence is reached.

 Step  2.1 Self-similarity regulation step:

  Step  2.1.1.  Use bicubic method to up scale the unrecovered LR patch 𝑦_LR to the same size 𝑛 as HR patch, defined as 𝑦_HR.

  Step  2.1.2.  Searching for a similar 𝑛 sized patch in 𝑦_LR’s neighbor:

   Step  2.1.2.1.  Compute each searching patch’s SSE as the self-similarity prior 𝐽(𝛼),

       𝐽(𝛼)=SSE=∑_𝑙∑_𝑘(𝐵_𝑙𝑘−𝐵_𝑙𝑘)²

   Step  2.1.2.2.  Find the least SSE patch, and compare its SSE with the adaptive threshold

        𝛼Var+𝛽. If SSE<𝛼Var+𝛽, define this least SSE patch as the similar patch 𝑆_HR.

  Step  2.1.3.  Use degradation matrix ℋ to down sample similar patch 𝑆_HR, define as 𝑆_LR.

  Step  2.1.4.  Subtract 𝑆_LR from LR patch 𝑦_LR, and get the residual 𝑅𝑆_LR=𝑦_LR−𝑆_LR.

  Step  2.1.5.  Recover the residual 𝑅𝑆_LR to 𝑅𝑆_HR using IRLS algorithm according (9).

  Step  2.1.6.  Add the 𝑅𝑆_HR to 𝑆_HR, according to (10).

 Step  2.2  Sparse dictionary regulation step: update 𝑥^𝑙+1_HR according to (11).

Step  3.  Ensemble all 𝑥_HR to recover HR image 𝑋 (if there is pixel overlap, the weighted average method is needed).