Add citations to doc strings

Author: constantinpape

src/bioimage_cpp/graph/lifted_multicut/__init__.py

@@ -322,6 +322,13 @@ def optimize(self, objective: LiftedMulticutObjective) -> np.ndarray:
 
 
 class LiftedGreedyAdditiveMulticut(LiftedMulticutSolver):
+    """Greedy additive edge contraction (GAEC) lifted multicut solver.
+
+    Introduced in "An efficient fusion move algorithm for the minimum cost
+    lifted multicut problem":
+    https://hci.iwr.uni-heidelberg.de/sites/default/files/publications/files/1939997197/beier_16_efficient.pdf
+    """
+
     def __init__(
         self,
         *,
@@ -362,6 +369,13 @@ def _build_cpp_sub_solver(self):
 
 
 class LiftedKernighanLinMulticut(LiftedMulticutSolver):
+    """Kernighan-Lin lifted multicut solver.
+
+    Introduced in "Efficient decomposition of image and mesh graphs by lifted
+    multicuts":
+    http://openaccess.thecvf.com/content_iccv_2015/papers/Keuper_Efficient_Decomposition_of_ICCV_2015_paper.pdf
+    """
+
     def __init__(
         self,
         *,
@@ -411,6 +425,10 @@ def _build_cpp_sub_solver(self):
 class FusionMoveLiftedMulticut(LiftedMulticutSolver):
     """Fusion-move lifted multicut solver.
 
+    Introduced in "An efficient fusion move algorithm for the minimum cost
+    lifted multicut problem":
+    https://hci.iwr.uni-heidelberg.de/sites/default/files/publications/files/1939997197/beier_16_efficient.pdf
+
     Iteratively generates proposals via ``proposal_generator`` (which sees the
     *base* graph and base edge costs), fuses them with the current best
     labeling, and accepts improvements. Each fuse contracts the base graph by

src/bioimage_cpp/graph/multicut/__init__.py

@@ -94,6 +94,12 @@ def optimize(self, objective: MulticutObjective) -> np.ndarray:
 
 
 class GreedyAdditiveMulticut(MulticutSolver):
+    """Greedy additive edge contraction (GAEC) multicut solver.
+
+    Introduced in "Fusion moves for correlation clustering":
+    http://openaccess.thecvf.com/content_cvpr_2015/papers/Beier_Fusion_Moves_for_2015_CVPR_paper.pdf
+    """
+
     def __init__(
         self,
         *,
@@ -133,6 +139,13 @@ def _build_cpp_sub_solver(self):
 
 
 class GreedyFixationMulticut(MulticutSolver):
+    """Greedy fixation multicut solver.
+
+    Introduced in "A Comparative Study of Local Search Algorithms for
+    Correlation Clustering":
+    https://link.springer.com/chapter/10.1007/978-3-319-66709-6_9
+    """
+
     def __init__(self, *, weight_stop: float = 0.0, node_num_stop: float = -1.0):
         self.weight_stop = float(weight_stop)
         self.node_num_stop = float(node_num_stop)
@@ -155,6 +168,12 @@ def _build_cpp_sub_solver(self):
 
 
 class KernighanLinMulticut(MulticutSolver):
+    """Kernighan-Lin multicut solver.
+
+    Introduced in "An efficient heuristic procedure for partitioning graphs":
+    http://xilinx.asia/_hdl/4/eda.ee.ucla.edu/EE201A-04Spring/kl.pdf
+    """
+
     def __init__(
         self,
         *,
@@ -289,6 +308,9 @@ def _build_for_thread(self, graph, edge_costs, seed_offset):
 class FusionMoveMulticut(MulticutSolver):
     """Fusion-move multicut solver.
 
+    Introduced in "Fusion moves for correlation clustering":
+    http://openaccess.thecvf.com/content_cvpr_2015/papers/Beier_Fusion_Moves_for_2015_CVPR_paper.pdf
+
     Iteratively generates proposals via ``proposal_generator``, fuses them
     with the current best labeling, and accepts improvements. The fuse step
     solves a contracted multicut subproblem with ``sub_solver``; if omitted,
@@ -396,6 +418,16 @@ def optimize(self, objective: MulticutObjective) -> np.ndarray:
 
 
 class MulticutDecomposer(MulticutSolver):
+    """Decomposition-based multicut solver.
+
+    Introduced in "Break and Conquer: Efficient Correlation Clustering for
+    Image Segmentation":
+    https://link.springer.com/chapter/10.1007/978-3-642-39140-8_9
+
+    Splits the multicut problem into connected components (based on positive
+    edge costs) and solves each component independently with ``sub_solver``.
+    """
+
     def __init__(
         self,
         sub_solver: MulticutSolver,

src/bioimage_cpp/graph/mutex_watershed.py

@@ -39,6 +39,10 @@ def mutex_watershed_clustering(
 ) -> np.ndarray:
     """Mutex watershed clustering on an undirected graph.
 
+    Introduced in "The Mutex Watershed and its Objective: Efficient,
+    Parameter-Free Image Partitioning":
+    https://arxiv.org/pdf/1904.12654.pdf
+
     Attractive edges come from ``graph`` (one cost per edge in
     ``edge_costs``); repulsive long-range edges are supplied separately as
     ``mutex_uvs`` with weights ``mutex_costs``. All edges are jointly sorted
@@ -113,6 +117,10 @@ def semantic_mutex_watershed_clustering(
 ) -> tuple[np.ndarray, np.ndarray]:
     """Semantic mutex watershed clustering on an undirected graph.
 
+    Introduced in "The Semantic Mutex Watershed for Efficient Bottom-Up
+    Semantic Instance Segmentation":
+    https://arxiv.org/pdf/1912.12717.pdf
+
     Extends :func:`mutex_watershed_clustering` with a third group of edges
     that attach semantic class labels to clusters. Two clusters carrying
     different semantic class labels are forbidden from merging; otherwise

src/bioimage_cpp/segmentation/mutex_watershed.py

@@ -30,6 +30,10 @@ def mutex_watershed(
 ) -> np.ndarray:
     """Run mutex watershed on a 2D or 3D image-derived grid graph.
 
+    Introduced in "The Mutex Watershed and its Objective: Efficient,
+    Parameter-Free Image Partitioning":
+    https://arxiv.org/pdf/1904.12654.pdf
+
     Parameters
     ----------
     affinities:
@@ -127,6 +131,10 @@ def semantic_mutex_watershed(
 ) -> tuple[np.ndarray, np.ndarray]:
     """Run semantic mutex watershed on a 2D or 3D image-derived grid graph.
 
+    Introduced in "The Semantic Mutex Watershed for Efficient Bottom-Up
+    Semantic Instance Segmentation":
+    https://arxiv.org/pdf/1912.12717.pdf
+
     The affinity array stacks three groups of channels along axis 0:
 
     1. ``[0, number_of_attractive_channels)`` — attractive grid edges.