Processing image data sets

1. A method, comprising:

applying a hierarchical bricking to an image data set that depicts semiconductor components, bricking comprising a plurality of bricks on a plurality of hierarchical levels, bricks on different hierarchical levels having different image element sizes of corresponding image elements; and

determining the hierarchical levels and/or edge lengths of the bricks on the plurality of hierarchical levels in accordance with structure sizes of the semiconductor components,

wherein:

the plurality of hierarchical levels comprise a first hierarchical level, a second hierarchical level, and a third hierarchical level;

each image element in the first hierarchical level has a first size;

each image element in the second hierarchical level has a second size;

each image element in the third hierarchical level has a third size;

a ratio of the first size to the second size has a first value;

a ratio of the second size to the third size has a second value;

the second value is different from the first value; and

one of the following holds:

i) for each of the image elements, the image element comprises a pixel and the corresponding image elements comprise corresponding pixels; or

ii) for each of the image elements, the image element comprises a voxel and the corresponding image elements comprise corresponding voxels.

2. The method of claim 1 , further comprising:

using an imaging device to acquire the image data set; and

obtaining the image data set from the imaging device.

3. The method of claim 1 , wherein the image data set comprises a three-dimensional image data set.

4. The method of claim 1 , further comprising determining the image element sizes and/or an image element number of the image elements of the bricks on the plurality of hierarchical levels in accordance with structure sizes of the semiconductor components.

5. The method of claim 1 , further comprising looking up the structure sizes of the semiconductor components in a database depending on a sample type of a sample depicted by the image data set.

6. The method of claim 1 , further comprising allocating labels to structures in the bricks on the plurality of hierarchical levels.

7. The method of claim 6 , wherein:

for structures in a brick on a lower hierarchical level of the bricking, a label is allocated on the basis of context information; and

the context information takes account of a predefined relationship between the structures corresponding to the semiconductor components in the brick on the lower hierarchical level and further structures corresponding to the semiconductor components in a further brick on an upper hierarchical level.

8. The method of claim 7 , wherein the predefined relationship is based on geometric reference data for the semiconductor components in the lower hierarchical level and in the upper hierarchical level.

9. The method of claim 6 , wherein:

the labels of the structures in bricks on a lower hierarchical level are selected from a first candidate set comprising geometric basic forms; and

the labels of the structures in bricks on an upper hierarchical level are selected from a second candidate set comprising a comprehensive overview of the semiconductor components, a logic group having a plurality of semiconductor components, interconnects, and/or lithography markers.

10. The method of claim 6 , wherein allocating the labels comprises, for structures in a brick on a lower hierarchical level, inheriting labels from one or more structures in a further brick on an upper hierarchical level, the further brick being associated with the brick on the lower hierarchical level, to obtain a multihierarchical label for the structures in the brick on in the lower hierarchical level.

11. The method of claim 6 , wherein allocating the labels comprises automatic or partially supervised propagation of labels between structures in bricks on a common hierarchical level.

12. The method of claim 11 , wherein propagating is effected based on geometric reference data for the semiconductor components in the bricks on the common level.

13. The method of claim 11 , wherein:

propagating is based on a machine-learned propagation algorithm; and

the method further comprises end-to-end training of the propagation algorithm or iterative online training of the propagation algorithm.

14. The method of claim 6 , wherein allocating the labels comprises:

determining label candidates for structures in associated bricks of different hierarchical levels via one or a plurality of first classification algorithms; and

determining the labels for the structures in the associated bricks via a second classification algorithm, which obtains the label candidates as input.

15. The method as claimed in claim 14 , wherein:

the second classification algorithm compares the label candidates of the structures in the associated bricks of different hierarchical levels with one another to obtain context information for the structures; and

the second classification algorithm plausibilizes the context information on the basis of a predefined relationship between structures on different hierarchical levels and determines the labels on the basis of the plausibilizing.

16. The method of claim 6 , wherein at least one of the following hold:

the labels classify the structures in at least one class;

the labels segment the structures in the image space; and

the labels provide the structures with a bounding box.

17. The method of claim 6 , further comprising controlling a human-machine interface to navigate through the image data set on the basis of the labels.

18. The method of claim 1 , further comprising performing an anomaly detection to recognize unknown structures in the bricks of different hierarchical levels.

19. The method of claim 1 , further comprising registering structures in the bricks on the plurality of hierarchical levels in each case with reference bricks of the corresponding hierarchical level.

20. The method of claim 1 , further comprising discarding at least some bricks on a hierarchical level depending on redundancies in depicted structures.

21. One or more machine-readable hardware storage devices comprising instructions that are executable by one or more processing devices to perform operations comprising the method of claim 1 .

22. A system comprising:

one or more processing devices; and

one or more machine-readable hardware storage devices comprising instructions that are executable by the one or more processing devices to perform operations comprising the method of claim 1 .

23. The method of claim 1 , wherein, for each of the image elements, the image element comprises a pixel, and the corresponding image elements comprise corresponding pixels.

24. The method of claim 1 , wherein, for each of the image elements, the image element comprises a voxel, and the corresponding image elements comprise corresponding voxels.