Ultrasound imaging device

Abstract

This ultrasound imaging apparatus has a learning result as a relationship between shape data and ultrasound incoming data which are obtained by performing deep learning using pedestal shape data (arrangement of multiple elements) and ultrasound incoming data of a target to be imaged obtained through transmission and reception of an ultrasound signal between multiple elements. The learning result is then applied to ultrasound incoming data to obtain shape estimation data as data on an estimated shape of a pedestal (estimated arrangement of the multiple elements), and image data on the target to be imaged are configured on the basis of the shape estimation data and the ultrasound incoming data.

Claims

1. An ultrasonic imaging apparatus comprising a probe having a base formed of a deformable plate-like material on which a plurality of elements capable of transmitting and / or receiving ultrasonic signals are arranged in an aligned manner, The system has a learning result as a relationship between the ultrasonic reception data and the shape data obtained by performing deep learning using the ultrasonic reception data of the imaging target obtained by transmitting and receiving the ultrasonic signal using the plurality of elements and the shape data of the base, The learning results are applied to the ultrasonic reception data to obtain estimated shape data as estimated shape data of the base, and an image of the target to be captured is constructed based on the estimated shape data and the ultrasonic reception data. An ultrasonic imaging device characterized by the following features.

2. An ultrasonic imaging apparatus according to claim 1, The estimated shape data is a sequence of coefficients for each basis function in a shape function that is represented by a linear combination of arbitrary basis function sequences. Ultrasound imaging device.

3. An ultrasonic imaging apparatus according to claim 2, The aforementioned shape function is P(1)sin(x) + P(2)sin(2x) + ... + P(n)sin(nx). Ultrasound imaging device.

4. An ultrasonic imaging apparatus according to claim 2 or 3, The deep learning method obtains the learning result by using the random shape data of the base obtained when the plurality of elements are randomly arranged on a processed image obtained by edge detection processing on a natural image, and the simulated received data obtained by sending and receiving the ultrasonic signal using the plurality of elements in the random shape data on the processed image through simulation, as the shape data and the ultrasonic received data, respectively. Ultrasound imaging device.

5. An ultrasonic imaging apparatus according to claim 2 or 3, The deep learning method obtains the learning result by using the measured shape data of the base and the measured received data from the probe, when the probe is attached to a living organism as the imaging target, as the shape data and the ultrasound received data, respectively. Ultrasound imaging device.

6. An ultrasonic imaging apparatus according to claim 4, The learning results are obtained by performing deep learning using the random shape data and the simulation reception data as shape data and ultrasound reception data, and further by performing additional deep learning using the measured shape data of the base and the measured reception data from the probe when the probe is attached to a living organism as the imaging target, as shape data and ultrasound reception data. Ultrasound imaging device.

Images