Sales of fresh duck fatty livers (foies gras) are mainly concentrated during the festive season, creating an imbalance between French consumer demand and product availability. As a result, some operators on the French market are tempted to freeze duck fatty livers in order to sell them fresh, a practice prohibited by French and European regulations. Slaughterers would like to be able to qualify products sold fresh, to ensure that they comply with regulations. To meet this objective, we developed a methodology for detecting fatty livers that have undergone a freezing-thawing process. In our study, 5 methods were tested: oxidation measurement , fingerprinting by MALDI-TOF (Matrix-Assisted Laser Dissociation ionization -Time_Of-Flight) mass spectrometry, near infrared spectroscopy (NIRS), Nuclear Magnetic Resonance (NMR), MRI and conductivity. In the first phase of the project, MALDI-TOF fingerprints were able to classify livers from the same batch, calibrated by weight, with internal validation and recognition percentages ranging from 93 % to 100 %. Similarly, the NIRS method showed that it was possible to develop calibrations with model determination coefficients (R²c) of 0.82, and cross-validation coefficients (R²cv) of 0.80. In the second phase of the project, variability was incorporated into the origin of the livers, in order to provide a model specific to freezing/thawing and usable on all-origin livers. The MALDI-TOF and NIRS methods were able to recognize frozen fatty livers, but with classification errors and too much variability to validate the first-phase model. A validation phase of these two methods on a larger pool of fatty livers, from different technical itineraries seems necessary to have robust and reliable prediction equations. As for NMR and MRI measurements, despite encouraging results in the first phase, they could not be repeated on fatty livers of various origins (different slaughterhouses) due to poor control of the cold chain between sampling and analysis.