Checklist

• I have searched for similar issues.
• For Python issues, I have tested with the latest development wheel.
• I have checked the release documentation and the latest documentation (for main branch).

Describe the issue

Slices of boolean mask tensors are not converted to indices correctly when using advanced indexing.
See PR #7340

Steps to reproduce the bug

Tensor a = Tensor::Zeros({20}, core::Int64, Device("CUDA:0"));
a.SetItem(core::TensorKey::Slice(0, 5, 1), Tensor::Init<int64_t>(1, a.GetDevice()));
a.SetItem(core::TensorKey::Slice(10, 15, 1), Tensor::Init<int64_t>(1, a.GetDevice()));
Tensor mask = a.Gt(0);

Tensor subRangeofMask = mask.Slice(0, 10, 20);
Tensor maskedFullRange = a.IndexGet({mask});
Tensor maskedSubRange = a.Slice(0, 10, 20).IndexGet({subRangeOfMask});

EXPECT_EQ(maskedFullRange.Slice(0, 10, 20), maskedSubRange); // this fails!

Error message

No response

Expected behavior

The boolean mask should be converted to the index list by using the correct data pointer (offset by potential slicing)

Open3D, Python and System information

- Operating system: Ubuntu 24.04
- Python version: Python 3.8
- Open3D version: latest main (fb7088ceebef38d54c575b47935e568979b50954)
- System architecture: x86
- Is this a remote workstation?: no
- How did you install Open3D?: build from source
- Compiler version (if built from source): gcc 13.3

Additional information

No response