Introduction - Data generated by Data Independent Acquisition mass spectrometry (DIA-MS) is used to both identify and quantify peptides within a biological sample. The output of DIA-MS is influenced by multiple sources of noise that may be due to impurities introduced during sample preparation, reagents used in the liquid chromatography column, or from particles in the atmosphere. Most of these impurities cause chemical noise; its signal is distributed over a wide range in the retention time, has constant in mass over charge (m/z) and usually has a single charge. Random noise is generated by the machine and depends on its operating conditions such as gain, temperature and age. Random noise is of high frequency and could occur at any point in m/z vs retention time space. In micro-channel plate detectors, the collision of an ion with the detector generates a signal. When a host of ions collide with the detector there could be an accumulation of residue at the detector which results in a shift in the baseline. Many attempts have been made to de-noise MS data. Most attempts try to de-noise spectra [1-8], yet important spatial information is lost when the MS data is summed over retention time. Other techniques [9, 10] perform de-noising of single ion chromatograms, and unlike [1-8] some of the multi-dimensional information of the MS data is explored by these methods. However, these are not strictly two-dimensional de-noising techniques. Method - Crane is a technique for de-noising raw MS data via two-dimensional undecimated wavelet transform inspired by image de-noising techniques. Preliminary results show that it is a promising technique for removing chemical and random noise. Ultimately, this technique should increase the accuracy and reduce false detection rates of peptide / fragment detection pipelines such as OpenSWATH.