System and method for learning-guided electron microscopy
A system and method is provided for rapidly collecting high quality images of a specimen through controlling a re-focusable beam of an electron microscope. An intelligent acquisition system instructs the electron microscope to perform an initial low-resolution scan of a sample. A low-resolution image of the sample is received by the intelligent acquisition system as scanned image information from the electron microscope. The intelligent acquisition system then determines regions of interest within the low-resolution image and instructs the electron microscope to perform a high-resolution scan of the sample, only in areas of the sample corresponding to the determined regions of interest or portions of the determined regions of interest, so that other regions within the sample are not scanned at high-resolution, where the high-resolution scanning in the regions of interest is guided by a probability map using a deep neural network for segmentation.
1. A method for rapidly collecting high quality images of a specimen through controlling a re-focusable beam of an electron microscope, wherein the method comprises the steps of:
instructing the electron microscope to perform an initial low-resolution scan of a sample;
receiving a low-resolution image of the sample as scanned image information from the electron microscope;
determining regions of interest within the low-resolution image; and
instructing the electron microscope to perform a high-resolution scan of the sample, only in areas of the sample corresponding to the determined regions of interest, so that other regions within the sample are not scanned at high-resolution, where the high-resolution scanning in the regions of interest is guided by a probability map using a deep neural network for segmentation.
2. The method of claim 1 , wherein a probability value of the probability map defines which pixel of the regions of interest is to be scanned.
3. The method of claim 2 , wherein probability values are used to define a pixel scanning order of the regions of interest.
4. The method of claim 3 , wherein pixels of the regions of interest with a larger probability value are first scanned, followed by pixels of the regions of interest having a lower probability value.
5. The method of claim 1 , further comprising the step of applying a reconstruction model to the received low resolution image, where the reconstruction model is supervised and regression-based (SRUNET).
6. The method of claim 5 , where the regression-based SRUNET is derived with supervised machine learning.
7. The method of claim 1 , wherein an initial scan percentage of the sample is 1/64 th of the sample.
8. The method of claim 1 , wherein an initial scan percentage of the sample is 1/16 th of the sample.
9. The method of claim 1 , further comprising the steps of:
applying a reconstruction model to the received low resolution image, resulting in a reconstructed image; and
comparing the reconstructed image to a target scanned high resolution image of the sample to determine a region of interest error.
10. The method of claim 1 , wherein the high-resolution scanning in the regions of interest is guided by a bitmap that is generated from the probability map using a deep neural network for segmentation.
11. A method for rapidly collecting high quality images of a specimen through controlling a re-focusable beam of an electron microscope, wherein the method comprises the steps of:
instructing the electron microscope to perform an initial low-resolution scan of a sample;
receiving a low-resolution image of the sample as scanned image information from the electron microscope; and
instructing the electron microscope to perform a high-resolution scan of the sample, only in areas of the sample corresponding to regions of interest, so that other regions within the sample are not scanned at high-resolution, where the high-resolution scanning in the regions of interest is guided by a bitmap that is generated from a probability map using a deep neural network for segmentation,
wherein the regions of interest are not defined and where an estimated reconstruction error is used to guide the high-resolution scan of the sample.
12. The method of claim 11 , wherein a probability value of the probability map defines which pixel of the regions of interest is to be scanned.
13. The method of claim 12 , wherein probability values are used to define a pixel scanning order of the regions of interest.
14. The method of claim 11 , further comprising the step of applying a reconstruction model to the received low resolution image, where the reconstruction model is supervised and regression-based (SRUNET).
15. The method of claim 14 , where the regression-based SRUNET is derived with supervised machine learning.