Enhanced thermographic analysis using complete laws' texture feature set for early cancer detection in veterinary medicine

This study presents a verification of a previously published method for detecting bone cancer in canines using thermographic imaging and texture analysis. The original work demonstrated high classification accuracy using a subset of Laws texture energy features. In our implementation, we conducted a systematic evaluation using the complete set of fifteen Laws features, made feasible by recent improvements to the FEPC tool that now support large-scale combinatoric testing. Beyond verification, the results reveal that a smaller subset of Laws features can reproduce the same classification performance while substantially reducing the computational burden associated with large feature sets. This observation introduces an opportunity for improved computational efficiency, which is particularly relevant for time-sensitive applications and resource-constrained environments. The findings contribute both a validated benchmark and a refined configuration for automated, non-invasive cancer detection in veterinary diagnostics.