The presence of shallow gas anomalies continues to be a challenge in seismic imaging in the Gulf of Mexico (GOM). Due to the strong absorption property and irregular shape of a gas cloud, the seismic image below it could suffer from frequency, amplitude, and phase distortion. Q tomography and pre-stack depth Q migration (Q-PSDM) are efficient tools for compensating these distortion effects caused by absorptive heterogeneities. However, deriving Q models is a challenging task, especially when the shape of gas anomalies are complex and the velocity inside or close to the gas cloud area is not correct. Standard ray-based tomography fails to capture the detailed rapid velocity variation for the areas with or beneath a high lateral or vertical contrast (e.g., gas pocket, shale, salt, and carbonate). Full waveform inversion (FWI) provides high-resolution velocity details but is still an expensive process with its success highly dependent on the quality of the initial model. We present a case study at East Breaks, GOM, that combines adaptive data-selection FWI and Q tomography to invert the velocity and absorption model around a gas cloud area to improve the seismic image.