One technique used to infer and monitor patient's respiratory conditions is the electrical impedance tomography (EIT). This provides images with information about lung function. The EIT image contrast is dependent on the variation of electrical impedance, therefore, this image does not provide anatomical details in border regions of several organs. To contribute to a clinical solution, we propose a new method to delimit regions of interest such as the pulmonary region and to improve the reconstruction quality of the EIT. Using a Matlab Toolbox k-wave, the ultrasound propagation phenomenon in homogeneous medium without patient (Reference) and with thoracic models were simulated, separately via a set of several ultrasound transducers distributed around the chest. After pulse emission by a transducer (TR), all received signals were compared considering the two sets of signals. If the energy relation between parts of the signals does not exceed an empirical threshold (30% in this study), a partial mask is generated between the transmitter and the receptor. This process was repeated until all 128 transducers are considered as TR-emitters. The 128 transducers (150kHz) are uniformly distributed. The evaluation was made by visually comparing the resulting images with the respective simulated object. A simple approach was presented to delimit high contrast organs with ultrasound transducers distributed around the patient. This approach allows other lower contrast objects to become invisible by varying the threshold limit. The investigation, based on numerical simulations of ultrasonic propagation, has shown promising results in the delimitation of the pulmonary region.