Tracking 2d

module demonstrating the usage of intensity based registration tracking 2d algorithm More...

module demonstrating the usage of intensity based registration tracking 2d algorithm

Overview

This script allows to track a rigid motion transform over a sequence of 2d images.

See also

intensity based registration tracker 2d algorithm

Usage

The application is a standalone script user can navigate through image plans of sequence using following keys :

• left key : go to previous plan
• right key : go to next plan
• up key : go to last plan
• down key : go to first plan

Note

this script requieres PyQt5 (see https://pypi.python.org/pypi/PyQt5 for example)

Here is a snapshot algorithm expected behavior:

Source code documentation

We start by importing all necessary libraries:

import sys
import os
import PyIPSDK
import PyIPSDK.IPSDKIPLFeatureDetection as fd
import PyIPSDK.IPSDKIPLRegistration as registration
from PyQt5.QtWidgets import QWidget, QApplication, QLabel
from PyQt5.QtGui import QPainter, QColor, QFont, QPen, QBrush, QImage, QPixmap, QPolygon
from PyQt5.QtCore import Qt, QCoreApplication, QPoint

Then we use tracking algorithm to compute motion registration over image sequence.

        # open sequence image
        imagesSamplePath = PyIPSDK.getIPSDKDirectory(PyIPSDK.eInternalDirectory.eID_Images)
        inputImgPath = os.path.join(imagesSamplePath,  "rotation_seq_2d_1.tif")
        self.seqImg = PyIPSDK.loadTiffImageFile(inputImgPath, PyIPSDK.eTiffDirectoryMode.eTDM_Temporal)
        firstImgPlan = PyIPSDK.extractPlan(0, 0, 0, self.seqImg)        
        
        # extract harris features
        nbSamples = 400
        minDist = 5
        pixels2d = fd.harrisCorner2d(firstImgPlan, nbSamples, minDist)
        
        # process registration
        self.initGrid = PyIPSDK.createRegistrationTracking2dGrid(self.seqImg.getSizeX() * 0.25,
                                                                 self.seqImg.getSizeY() * 0.25,
                                                                 self.seqImg.getSizeX() * 0.75,
                                                                 self.seqImg.getSizeY() * 0.75)
        self.initPts = PyIPSDK.toCoords2dColl(pixels2d)
        self.piMotionTransform2d = registration.intensityBasedTracker2d(self.seqImg, self.initPts,
                                                                        PyIPSDK.eRegistrationMotionModel2d.eRMM2d_Rigid,
                                                                        self.initGrid)

See the full source listing