Medical imaging device, image processing method, and program
In a medical imaging device that performs compressed sensing, it is possible to shorten a reconstruction time while maintaining image quality. The medical imaging device includes an image reconstructing unit that reconstructs an image by performing an iterative optimization operation of compressed sensing and a base selecting unit that selects a base transform which is used for the optimization every iteration. The base selecting unit may select a base on the basis of a predetermined base sequence or may select a base using weighting factors which are set for the bases in advance. The invention is applied to a medical imaging device such as an MRI apparatus, an ultrasonic imaging apparatus, or a CT apparatus.
1. A medical imaging device comprising:
an imaging unit configured to collect data which is required for image reconstruction from an examination object; and
a data processing unit configured to process data which is collected by the imaging unit,
wherein the data processing unit includes an image reconstructing unit configured to perform an iterative operation on sparse data collected by the imaging unit by using a plurality of base transforms to reconstruct an image and a base selecting unit configured to select a base transform whose number is smaller than the plurality of base transforms which are used for the iterative operation,
the base transform which is selected by the base selecting unit differs in at least two iterations, and
the image reconstructing unit performs the iterative operation using the base transform which is selected by the base selecting unit.
2. The medical imaging device according to claim 1 , wherein the base selecting unit selects the base transform in accordance with a predetermined sequence of the plurality of base transforms.
3. The medical imaging device according to claim 1 , wherein the base selecting unit selects the base transform on the basis of at least one of a processing time using the base transform and image quality of an image which is acquired after the base transform has been performed.
4. The medical imaging device according to claim 1 , further comprising a weight calculating unit configured to calculate weights for the plurality of base transforms,
wherein the base selecting unit selects the base transform using the weights calculated by the weight calculating unit.
5. The medical imaging device according to claim 4 , further comprising an input unit that is used for a user to input an index for selecting the base transform,
wherein the weight calculating unit calculates the weights on the basis of the index input from the input unit.
6. The medical imaging device according to claim 4 , wherein the weight calculating unit calculates the weights on the basis of imaging conditions which are used for the imaging unit to acquire the data.
7. The medical imaging device according to claim 6 , wherein the base selecting unit prepares a selection sequence table in which a sequence for selecting the base transform is defined using the weights calculated by the weight calculating unit and selects the base transform on the basis of the selection sequence table.
8. The medical imaging device according to claim 1 , wherein the medical imaging device is one of a magnetic resonance imaging apparatus, an ultrasonic imaging apparatus, and a computed tomography (CT) apparatus.
9. An image processing method of reconstructing an image by performing an iterative optimization operation using a plurality of base transforms on sparse observed data which is acquired by an imaging unit of a medical imaging device, the image processing method comprising:
selecting one base transform or a smaller number of base transforms than the plurality of base transforms from the plurality of base transforms; and
performing an iterative operation using the base transform which differs in at least two iterations.
10. The image processing method according to claim 9 , wherein the base transform to be selected is changed every iteration.
11. The image processing method according to claim 9 , further comprising a step of calculating weights for the plurality of base transforms,
wherein the base transform for each iteration is selected on the basis of the weights.
12. The image processing method according to claim 9 , wherein the base transform which is selected every iteration is one base transform and a plurality of base transforms are used in the iterative operation as a whole.
13. A program causing a computer to perform following steps A to E which are:
Step A of calculating estimated data with a minimized error using sparse observed data, which is collected by a medical imaging device, as an initial value;
Step B of selecting a smaller number of base transforms than a plurality of base transform candidates form the candidates;
Step C of performing an L1 norm minimization on the estimated data calculated in the error minimization step A using the selected base transforms;
Step D of updating the initial value in the error minimizing step using a coefficient acquired by the L1 norm minimization; and
Step E of repeating Steps A to D until ending conditions are satisfied.