RMSD_Transformation_Input/Output_Data

This repository contains synthetic and real seismic SEG-Y datasets generated for the development, evaluation, and reproducibility of the RMSD-based seismic waveform transformation framework. The archive includes original seismic sections together with deterministic rule-transformed, machine-learning-transformed, and ML-proxy-transformed seismic datasets. The RMSD framework operates through localized inter-peak waveform conditioning, selectively modifying short positive runs while preserving the dominant reflection peaks associated with principal impedance contrasts. Unlike globally acting filtering or deconvolution methods, the proposed approach performs sparse waveform transformation designed to enhance reflector continuity while minimizing large-scale spectral distortion. The repository includes: Synthetic 15 ms reflector-separation SEG-Y datasets RMSD rule-transformed synthetic datasets Original WS70 seismic dataset RMSD rule-transformed WS70 seismic dataset RMSD ML-transformed WS70 seismic dataset RMSD ML-proxy transformed WS70 seismic dataset Supporting metadata and README documentation The datasets are intended for: seismic waveform-conditioning research, reflector continuity enhancement studies, sparse polarity transformation analysis, benchmarking against conventional seismic enhancement methods, ML-assisted seismic transformation workflows, and reproducible geophysical research. The repository supports investigations into localized waveform conditioning, thin-bed interference behavior, feature-engineered seismic transformation, and scalable ML-assisted seismic processing. Data format:IEEE floating-point SEG-Y