arXiv:2507.06325v3 Announce Type: replace Abstract: Fractal image compression (FIC) can achieve high compression ratios and fast decoding, but its encoding stage requires an expensive search for matching image blocks. This paper evaluates two implementation strategies for reducing this cost: isometry pruning, and fractal-dimension-based search gating. The gating method uses a box-counting estimate of range-block complexity to determine whether the domain-pool search should be performed. Experiments on the Monkey, Einstein, and Cameraman grayscale images examine two domain/range block-size configurations. The results show that isometry pruning substantially reduces total execution time. Combining isometry pruning with box-counting gating further reduces the number of searched blocks and the bitrate, although the resulting reduction in PSNR depends on the selected thresholds. These findings characterize the trade-off among execution time, bitrate, and reconstruction quality.