feat: Runtime optimizations for MultiLabelConfusionMatrix

[nikos-livathinos]

Improve the runtime execution of the pixel-wise layout evaluations by compressing the binary representations given in the input of the MultiLabelConfusionMatrix.compute_confusion_matrix().

The idea is:

1. Get element-wise pairs from the binary representations of gt and preds.
2. Find the unique pairs and the counts of each pair.
3. Compute the contributions of each unique binary pair to the final confusion matrix (no change to the prior implementation).
4. Multiply 3 by the weight (count) of the pair.

This compression technique significantly reduces the computational load, since most GT–prediction pairs repeat.

[github-actions]

✅ DCO Check Passed

Thanks @nikos-livathinos, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

Your pull request matches the following merge protections and will not be merged until they are valid.

🟢 Enforce conventional commit

Wonderful, this rule succeeded.

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?: