Implements road finding using GNNs for pattern recognition. More...

#include "AthenaBaseComps/AthAlgTool.h"
#include "GaudiKernel/ToolHandle.h"
#include "FPGATrackSimHough/IFPGATrackSimRoadFinderTool.h"
#include "FPGATrackSimGNNGraphHitSelectorTool.h"
#include "FPGATrackSimGNNGraphConstructionTool.h"
#include "FPGATrackSimGNNEdgeClassifierTool.h"
#include "FPGATrackSimGNNRoadMakerTool.h"
#include "FPGATrackSimGNNRootOutputTool.h"

Include dependency graph for FPGATrackSimGNNPatternRecoTool.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	FPGATrackSimGNNPatternRecoTool

Detailed Description

Implements road finding using GNNs for pattern recognition.

Author: Jared Burleson - jared.nosp@m..dyn.nosp@m.es.bu.nosp@m.rles.nosp@m.on@ce.nosp@m.rn.c.nosp@m.h

Date: November 27th, 2024 This class serves as the overhead for using GNNs for pattern recognition in FPGATrackSim. The getRoads() function from IFPGATrackSimRoadFinderTool is declared here and it will use the tools that are included in this file. Each tool will be part of the pipeline necessary to complete getRoads().

The GNN Pattern Recognition Tool has four phases:

Graph Hit Selector Tool : select hits needed for pattern recognition and record additional information
Graph Construction Tool : use hits to build edges necessary for the GNN, which are spatial connections between hits
Edge Classifier Tool : run inference of Interaction Network GNN model to assign a score to each edge
Road Maker Tool : use edge scores to select hits that can be algorithmically identified in a path, called a road

The output of this tool will feed back into the FPGATrackSim Analysis pipeline. Currently this code does not operate with subRegions(), but only using the region defined by the map defined in the input python script for hit selection. This is handled in the data preparation algorithm before this code.

Definition in file FPGATrackSimGNNPatternRecoTool.h.

Classes

Detailed Description