python-microscopy 25.11.20__cp313-cp313-win_amd64.whl

PYME/DSView/modules/blobMeasure.py

@@ -0,0 +1,1093 @@
+#!/usr/bin/python
+
+##################
+# blobMeasure.py
+#
+# Copyright David Baddeley, 2012
+# d.baddeley@auckland.ac.nz
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+##################
+import numpy as np
+from numpy import linalg
+from scipy import ndimage
+import wx
+import wx.lib.agw.aui as aui
+
+from ._base import Plugin
+
+#TODO - move non-GUI logic out of here!!
+
+def iGen():
+    i = 1
+    while True:
+        i += 1
+        yield i
+        
+ig = iGen()
+
+def getNewID():
+    return next(ig)
+
+
+class DataBlock(object):
+    def __init__(self, data, X, Y, Z, voxelsize=(1,1,1)):
+        self.data = data
+        self.X = X
+        self.Y = Y
+        self.Z = Z
+        
+        self.voxelsize = voxelsize
+        
+    def __getstate__(self):
+        #support for pickling
+        return {'data':self.data, 'X':self.X, 'Y':self.Y, 'Z':self.Z, 'voxelsize':self.voxelsize}
+        
+    @property
+    def bbox(self):
+        if not '_bbox' in dir(self):
+            x0 = self.X.min()*self.voxelsize[0]
+            y0 = self.Y.min()*self.voxelsize[1]
+            z0 = self.Z.min()*self.voxelsize[2]
+            
+            x1 = self.X.max()*self.voxelsize[0]
+            y1 = self.Y.max()*self.voxelsize[1]
+            z1 = self.Z.max()*self.voxelsize[2]
+            
+            self._bbox = (x0,y0,z0, x1, y1, z1)
+        return self._bbox
+    
+    @property
+    def sum(self):
+        if not '_sum' in dir(self):
+            self._sum = self.data.sum()
+        return self._sum
+    
+    @property    
+    def centroid(self):
+        if not '_centroid' in dir(self):
+            #crop off the bottom of the data to avoid biasing the centroid
+            m = np.maximum(self.data - 1.0*self.data.mean(), 0)
+            m = m/m.sum()
+            xc = (m*self.X).sum()
+            yc = (m*self.Y).sum()
+            zc = (m*self.Z).sum()
+    
+            self._centroid =  (xc, yc, zc)
+        
+        return self._centroid
+        
+    @property
+    def centroidNM(self):
+        return self.voxelsize[0]*self.centroid[0], self.voxelsize[1]*self.centroid[1], self.voxelsize[2]*self.centroid[2]
+        
+    @property
+    def XC(self):
+        if not '_XC' in dir(self):
+            self._XC = self.voxelsize[0]*(self.X - self.centroid[0])
+        return self._XC
+        
+    @property
+    def YC(self):
+        if not '_YC' in dir(self):
+            self._YC = self.voxelsize[1]*(self.Y - self.centroid[1])
+        return self._YC
+        
+    @property
+    def ZC(self):
+        if not '_ZC' in dir(self):
+            self._ZC = self.voxelsize[2]*(self.Z - self.centroid[2])
+        return self._ZC
+        
+    @property
+    def principalAxis(self):
+        if not '_principalAxis' in dir(self):
+            #b = (self.XC*self.data).ravel()
+            b = np.vstack([(self.XC*self.data).ravel(), (self.YC*self.data).ravel(), (self.ZC*self.data).ravel()]).T
+            
+            #use thresholded data for determining principle axis
+            m = self.data
+            m = m > 0.2*m.max()
+            #print((m.sum()))
+            l, nl = ndimage.label(m)
+            
+            #take only the largest contiguous region
+            m = 0*m
+            for i in range(1,nl+1):
+                r = l == i
+                if r.sum() > m.sum():
+                    m = r
+            b = b[m.ravel()>0.5, :]
+            #print((m.sum(), b.shape))
+            if b.shape[0] < 2:
+                self._principalAxis = np.NaN*np.ones(3)
+            #print b.shape
+            #A = (self.R*self.data*np.sign(self.XC)).ravel()[:,None]
+            #print A.shape
+            #pz = (np.abs(self.ZC)*self.data).sum()
+            #px = 1.0
+            else:
+                pa = linalg.svd(b, full_matrices=False)[2][0]
+            #print pa
+            #pa = linalg.lstsq(A, b)[0].squeeze()
+            #print pa
+            
+            #print pa.shape
+            
+            #pa = np.array([1.0, (1 - /py, 1./pz])
+                self._principalAxis = pa/linalg.norm(pa)
+            
+        return self._principalAxis
+        
+    @property
+    def secondaryAxis(self):
+        if not '_secondaryAxis' in dir(self):
+            trial = np.roll(self.principalAxis, 1) #gauranteed not to be co-linear with pa
+            
+            sa = np.cross(self.principalAxis, trial)
+            #set the secondary and 3ary axes for now - will be updated later
+            sa /= linalg.norm(sa)
+            self._secondaryAxis = sa
+            
+            
+            b = np.vstack([(self.A1*self.data).ravel(), (self.A2*self.data).ravel()]).T
+            b = b[self.data.ravel()>0]
+            if b.shape[0] < 2:
+                self._secondaryAxis = np.NaN*np.ones(3)
+            else:
+                p1, p2 = linalg.svd(b, full_matrices=False)[2][0]
+                #p1 = (np.abs(self.A1)*self.data).sum()
+                #p2 = (np.abs(self.A2)*self.data).sum()
+            
+                sa = p1*sa + p2*self.tertiaryAxis
+                if self.data.shape[2] == 1:
+                    #2D, force sa to be in plane
+                    sa[2] = 0
+                sa /= linalg.norm(sa)
+                self._secondaryAxis = sa
+            
+            del self._tertiaryAxis, self._A1, self._A2 #force these to be re-calculated
+            
+        return self._secondaryAxis
+        
+    def flipSecondaryAxis(self):
+        self._secondaryAxis = -self._secondaryAxis
+        try:
+            del self._tertiaryAxis
+        except AttributeError:
+            pass
+        try:
+            del self._A1
+        except AttributeError:
+            pass
+        try:
+            del self._A2
+        except AttributeError:
+            pass
+        
+    @property
+    def tertiaryAxis(self):
+        if not '_tertiaryAxis' in dir(self):
+            ta = np.cross(self.principalAxis, self.secondaryAxis)
+            ta /= linalg.norm(ta)
+            self._tertiaryAxis = ta
+            
+        return self._tertiaryAxis
+    
+    
+    @property
+    def A0(self):
+        if not '_A0' in dir(self):
+            self._A0 = self.principalAxis[0]*self.XC + self.principalAxis[1]*self.YC + self.principalAxis[2]*self.ZC 
+        return self._A0
+        
+    @property
+    def A1(self):
+        if not '_A1' in dir(self):
+            self._A1 = self.secondaryAxis[0]*self.XC + self.secondaryAxis[1]*self.YC + self.secondaryAxis[2]*self.ZC 
+        return self._A1
+        
+    @property
+    def A2(self):
+        if not '_A2' in dir(self):
+            self._A2 = self.tertiaryAxis[0]*self.XC + self.tertiaryAxis[1]*self.YC + self.tertiaryAxis[2]*self.ZC 
+        return self._A2
+        
+    @property
+    def SA(self):
+        if self.data.ndim > 2 and self.data.shape[2] > 1:
+            return self.A2
+        else:
+            return self.A1
+            
+    @property
+    def R(self):
+        return np.sqrt(self.XC**2 + self.YC**2 + self.ZC**2)
+        
+    @property
+    def r(self):
+        return np.sqrt(self.A0**2 + self.A1**2)
+        
+    @property
+    def Theta(self):
+        return np.angle(self.A0 +1j*self.A1)
+        
+    @property
+    def Phi(self):
+        return np.angle(self.A1 +1j*self.A2)
+    
+    @property
+    def geom_length(self):
+        vals = self.A0[self.data>0]
+        if self.sum == 0:
+            return 0
+        else:
+            return vals.max() - vals.min()
+        
+    @property
+    def geom_width(self):
+        if self.sum == 0:
+            return 0
+        else:
+            vals = self.A1[self.data>0]
+            return vals.max() - vals.min()
+        
+    @property
+    def geom_depth(self):
+        if self.sum == 0:
+            return 0
+        else:
+            vals = self.A2[self.data>0]
+            return vals.max() - vals.min()
+        
+    @property
+    def mad_0(self):
+        return (abs(self.A0)*self.data).sum()/self.sum
+    
+    @property
+    def mad_1(self):
+        return (abs(self.A1)*self.data).sum()/self.sum
+    
+    @property
+    def mad_2(self):
+        return (abs(self.A2)*self.data).sum()/self.sum
+
+class BlobObject(object):
+    def __init__(self, channels, masterChan=0, orientChan=1, orient_dir=-1):
+        self.chans = channels
+        self.masterChan = masterChan
+        self.orientChan = orientChan
+        self.orient_dir = orient_dir
+        
+        self.shown = True
+        
+        self.nsteps = 51
+        if self.chans[0].voxelsize[0] > 20:
+            self.nsteps = 10
+            
+        mc = self.chans[masterChan]
+        oc = self.chans[orientChan]
+        
+        if (mc.SA*oc.data).sum()*orient_dir < 0:
+            #flip the direction of short axis on mc
+            mc.flipSecondaryAxis()
+            
+    def __getstate__(self):
+        return {'chans': self.chans, 'masterChan':self.masterChan, 
+                'orientChan':self.orientChan, 'orient_dir':self.orient_dir, 
+                'shown':self.shown, 'nsteps':self.nsteps}
+        
+    def __setstate__(self, state):
+        self.__dict__.update(state)
+        
+        mc = self.chans[self.masterChan]
+        oc = self.chans[self.orientChan]
+        
+        if (mc.SA*oc.data).sum()*self.orient_dir < 0:
+            #flip the direction of short axis on mc
+            mc.flipSecondaryAxis()
+        
+        
+    def longAxisDistN(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(-1,1, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(mc.A0/mc.geom_length, c.data, xvs)
+            bb = bm*bn
+            bms.append(bb/bb.sum())
+            
+        return xvs, bms
+    longAxisDistN.xlabel = 'Longitudinal position [a.u.]'
+    
+    def longAxisOrthDist(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(-1,1, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(mc.A0/mc.geom_length, c.data, xvs)
+            bss = bm*bn
+            bn, bm, bs = binAvg(mc.A0/mc.geom_length, c.data*mc.SA, xvs)
+            bb = bm*bn
+            bms.append(bb/bss)
+            
+        return xvs, bms
+    longAxisOrthDist.xlabel = 'Longitudinal position [a.u.]'
+        
+    def shortAxisDist(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(-500,500, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(mc.SA, c.data, xvs)
+            bb = bm*bn
+            bms.append(bb/bb.sum())
+            
+        return xvs, bms
+    shortAxisDist.xlabel = 'Position on shortest axis [nm]'
+        
+    def radialDistN(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(0, 2, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        r = mc.r/(mc.geom_length/2)
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(r, c.data, xvs)
+            bb = bm*bn
+            bms.append(bb/bb.sum())
+            
+        return xvs, bms
+    radialDistN.xlabel = 'Radial position [a.u.]'
+    
+    def RadialDistN(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(0, 1, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        R = mc.R/(mc.geom_length/2)
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(R, c.data, xvs)
+            bb = bm*bn
+            bms.append(bb/bb.sum())
+            
+        return xvs, bms
+    RadialDistN.xlabel = 'Radial position [a.u.]'
+        
+    def angularDist(self):
+        from PYME.Analysis.binAvg import binAvg
+        
+        xvs = np.linspace(-np.pi, np.pi, self.nsteps)
+        bms = []
+        
+        mc = self.chans[self.masterChan]
+        
+        for c in self.chans:
+            bn, bm, bs = binAvg(mc.Theta, c.data, xvs)
+            bb = bm*bn
+            bms.append(bb/bb.sum())
+            
+        return xvs, bms
+    angularDist.xlabel=None
+        
+    def drawOverlay(self, view, dc):
+        #import wx
+        if self.shown:
+            try:
+                mc = self.chans[self.masterChan]
+                
+                #x, y, z = mc.centroid
+                l = mc.geom_length/(2*mc.voxelsize[0])
+                x0, y0, z0 = mc.centroid - l*mc.principalAxis
+                x1, y1, z1 = mc.centroid + l*mc.principalAxis
+                
+                x0_, y0_ = view.pixel_to_screen_coordinates(x0, y0)
+                x1_, y1_ = view.pixel_to_screen_coordinates(x1, y1)
+                
+                dc.DrawLine(int(x0_), int(y0_), int(x1_), int(y1_))
+                
+                l = mc.geom_width/(2*mc.voxelsize[0])
+                x0, y0, z0 = mc.centroid
+                x1, y1, z1 = mc.centroid + l*mc.secondaryAxis
+                
+                x0_, y0_ = view.pixel_to_screen_coordinates(x0, y0)
+                x1_, y1_ = view.pixel_to_screen_coordinates(x1, y1)
+                
+                dc.DrawLine(int(x0_), int(y0_), int(x1_), int(y1_))
+            except ValueError:
+                pass
+            
+    def getImage(self, gains=None):
+        try:
+            import Image
+        except ImportError:
+            from PIL import Image
+            
+        from io import BytesIO
+        import cherrypy
+        cherrypy.response.headers["Content-Type"]="image/png"
+        
+        dispChans = []
+        ovsc = 2
+        if self.chans[0].voxelsize[0] > 50:
+            ovsc = 1
+            
+        for i, c in enumerate(self.chans):
+            d = c.data
+            if d.ndim > 2 and d.shape[2] > 1:
+                d = d.max(2)
+            d = np.atleast_3d(d.squeeze().T)
+            d -= d.min()
+            if gains is None:
+                d = np.minimum(ovsc*255*d/d.max(), 255).astype('uint8')
+            else:
+                d = np.minimum(gains[i]*255*d, 255).astype('uint8')
+            
+            dispChans.append(d)
+            
+        while len(dispChans) < 3:
+            dispChans.append(0*d)
+            
+        im = np.concatenate(dispChans, 2)
+        
+        #scalebar (200 nm)
+        nPixels = 200/(c.voxelsize[0])
+        w = max(int(15/c.voxelsize[0]), 1)
+        if c.voxelsize[0] < 50:
+            im[-(2*w):-w, -(nPixels+w):-w, :] = 255
+    
+
+        xsize = im.shape[0]
+        ysize = im.shape[1]
+
+        zoom = 200./max(xsize, ysize)
+        
+        out = BytesIO()
+        Image.fromarray(im).resize((int(zoom*ysize), int(zoom*xsize))).save(out, 'PNG')
+        s = out.getvalue()
+        out.close()
+        return s
+        
+    def getGraph(self, graphName):
+        #import Image
+        from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
+        from matplotlib.figure import Figure
+        from io import BytesIO
+        import cherrypy
+        cherrypy.response.headers["Content-Type"]="image/png"
+        
+        isPolar =  'ang' in graphName
+        if isPolar:
+            fig = Figure(figsize=(3,3), facecolor='none')
+        else:
+            fig = Figure(figsize=(4,3), facecolor='none')
+        canvas = FigureCanvas(fig)
+        ax = fig.add_axes([.1, .15, .85, .8], polar=isPolar)
+        
+        xv, yv = getattr(self, graphName)()
+        
+        cols = ['r','g', 'b']
+        for i in range(len(yv)):
+            ax.plot(xv[:-1], yv[i], c = cols[i], lw=2)
+        ax.set_xlabel(getattr(self, graphName).xlabel)
+        
+        out = BytesIO()
+        canvas.print_png(out, dpi=100, facecolor='w')
+        s = out.getvalue()
+        out.close()
+        return s
+        
+    def getSchematic(self):
+        #import Image
+        from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
+        from matplotlib.figure import Figure
+        from matplotlib.patches import Rectangle
+        from io import BytesIO
+        import cherrypy
+        cherrypy.response.headers["Content-Type"]="image/png"
+
+        fig = Figure(figsize=(2,2), facecolor='none')
+        canvas = FigureCanvas(fig)
+        ax = fig.add_axes([0.01, 0.01, .98, .98])
+        
+        mc = self.chans[self.masterChan]
+        
+        bx0, by0, bz0, bx1, by1, bz1 = mc.bbox
+        rect = Rectangle((bx0, by0), (bx1 - bx0), (by1 - by0), facecolor='w')
+        ax.add_patch(rect)
+            
+        #x, y, z = mc.centroid
+        l = mc.geom_length/2
+        x0, y0, z0 = mc.centroidNM - l*mc.principalAxis
+        x1, y1, z1 = mc.centroidNM + l*mc.principalAxis
+        
+        ax.plot([x0, x1], [y0, y1], 'k', lw=2)
+        
+        l = mc.geom_width/2
+        x0, y0, z0 = mc.centroidNM
+        x1, y1, z1 = mc.centroidNM + l*mc.secondaryAxis
+        ax.plot([x0, x1], [y0, y1], 'm', lw=2)
+        
+        l = mc.geom_depth/2
+        x0, y0, z0 = mc.centroidNM
+        x1, y1, z1 = mc.centroidNM + l*mc.tertiaryAxis
+        ax.plot([x0, x1], [y0, y1], 'c--', lw=2)
+        
+        cols = ['r','g', 'b']
+        for i, c in enumerate(self.chans):
+            l = c.geom_length/2
+            x0, y0, z0 = c.centroidNM - l*c.principalAxis
+            x1, y1, z1 = c.centroidNM + l*c.principalAxis
+            
+            ax.plot([x0, x1], [y0, y1],c = cols[i] , lw=2)
+            
+            x0, y0, z0 = c.centroidNM
+            ax.plot(x0, y0, 'x', c = cols[i], lw=2)
+            
+        
+        ax.axis('scaled')
+        ax.set_xlim(bx0, bx1)
+        ax.set_ylim(by1, by0)
+        
+        ax.set_axis_off()
+        
+        out = BytesIO()
+        canvas.print_png(out, dpi=100, facecolor='none')
+        s = out.getvalue()
+        out.close()
+        return s
+        
+    def getSchematic3D(self):
+        #import Image
+        from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
+        from mpl_toolkits.mplot3d import Axes3D
+        from matplotlib.figure import Figure
+        from io import BytesIO
+        import cherrypy
+        cherrypy.response.headers["Content-Type"]="image/png"
+
+        fig = Figure(figsize=(2,2), facecolor='none')
+        canvas = FigureCanvas(fig)
+        #ax = fig.add_axes([.1, .15, .85, .8])
+        ax = Axes3D(fig)
+        
+        mc = self.chans[self.masterChan]
+            
+        #x, y, z = mc.centroid
+        l = mc.geom_length/(2*mc.voxelsize[0])
+        x0, y0, z0 = mc.centroid - l*mc.principalAxis
+        x1, y1, z1 = mc.centroid + l*mc.principalAxis
+        
+        ax.plot([x0, x1], [y0, y1], [z0, z1], 'k', lw=2)
+        
+        l = mc.geom_width/(2*mc.voxelsize[0])
+        x0, y0, z0 = mc.centroid
+        x1, y1, z1 = mc.centroid + l*mc.secondaryAxis
+        ax.plot([x0, x1], [y0, y1], [z0, z1], 'k', lw=2)
+        
+        l = mc.geom_depth/(2*mc.voxelsize[0])
+        x0, y0, z0 = mc.centroid
+        x1, y1, z1 = mc.centroid + l*mc.tertiaryAxis
+        ax.plot([x0, x1], [y0, y1], [z0, z1], 'k:', lw=2)
+        
+        cols = ['r','g', 'b']
+        for i, c in enumerate(self.chans):
+            l = c.geom_length/(2*mc.voxelsize[0])
+            x0, y0, z0 = c.centroid - l*c.principalAxis
+            x1, y1, z1 = c.centroid + l*c.principalAxis
+            
+            ax.plot([x0, x1], [y0, y1], [z0, z1],c = cols[i] , lw=2)
+            
+            x0, y0, z0 = c.centroid
+            ax.plot([x0], [y0], [z0], 'x', c = cols[i], lw=2)
+            
+        #ax.axis('equal')
+        #ax.set_axis_off()
+        
+        out = BytesIO()
+        canvas.print_png(out, dpi=100, facecolor='none')
+        s = out.getvalue()
+        out.close()
+        return s
+        
+    def _shade(self, faces, baseColor= [1,0,0, 0], lightVector=[1,1,1]):
+        baseColor = np.array(baseColor)
+        lightVector = np.array(lightVector)
+        v1 = faces[:,1, :] - faces[:,0, :]
+        v2 = faces[:,2, :] - faces[:,1, :]
+        
+        vnorm = np.cross(v1, v2)
+        
+        vnorm /= np.linalg.norm(vnorm, axis=-1)[:,None]
+        
+        return baseColor[None, :]*np.clip(np.abs(np.inner(vnorm, lightVector)), .05, 1)[:,None]
+        
+        
+        
+    def get3DIsosurf(self, isovalues=None):
+        from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
+        from mpl_toolkits.mplot3d import Axes3D
+        from mpl_toolkits.mplot3d.art3d import Poly3DCollection
+        from matplotlib.figure import Figure
+        from skimage import measure
+        from io import BytesIO
+        import cherrypy
+        cherrypy.response.headers["Content-Type"]="image/png"
+        
+        if '_3diso' in dir(self):
+            return self._3diso
+
+        fig = Figure(figsize=(8,4), facecolor='none')
+        canvas = FigureCanvas(fig)
+        #ax = fig.add_axes([.1, .15, .85, .8])
+        ax = Axes3D(fig, [0, 0, .5,1], azim = -60, elev=30, axisbg='none')
+        #ax = fig.add_subplot(1, 2, 1, projection='3d')
+        
+        mc = self.chans[self.masterChan]
+
+        meshes = []        
+        cols = [[1.,0.,0.], [0.,1.,0.], [0.,0.,1.]]    
+        
+        for i, c in enumerate(self.chans):
+            d = c.data#.swapaxes(0,1)
+            verts, faces = measure.marching_cubes(d, isovalues[i])
+            vertsc = verts + np.array(c.bbox[:3])/np.array(mc.voxelsize)
+            tris =  vertsc[faces]
+            tricols = self._shade(tris, cols[i])
+            #mesh = Poly3DCollection(vertsc[faces], facecolor=['r', 'g', 'b'][i], edgecolor=['r', 'g', 'b'][i], alpha=.5)
+            #mesh = Poly3DCollection(tris, facecolor=self._shade(tris, cols[i]), edgecolor=['r', 'g', 'b'][i], alpha=.5)
+            meshes.append((tris, tricols))
+        
+        
+            
+        #x, y, z = mc.centroid
+            
+        def _drawaxes(ax):
+            l = mc.geom_length/(2*mc.voxelsize[0])
+            x0, y0, z0 = mc.centroid - l*mc.principalAxis
+            x1, y1, z1 = mc.centroid + l*mc.principalAxis
+            
+            ax.plot([x0, x1], [y0, y1], [z0, z1], 'k', lw=3)
+            
+            l = mc.geom_width/(2*mc.voxelsize[0])
+            x0, y0, z0 = mc.centroid
+            x1, y1, z1 = mc.centroid + l*mc.secondaryAxis
+            ax.plot([x0, x1], [y0, y1], [z0, z1], 'm', lw=3)
+            
+            l = mc.geom_depth/(2*mc.voxelsize[0])
+            x0, y0, z0 = mc.centroid
+            x1, y1, z1 = mc.centroid + l*mc.tertiaryAxis
+            ax.plot([x0, x1], [y0, y1], [z0, z1], 'c--', lw=3)
+            
+            cols = ['r','g', 'b']
+            for i, c in enumerate(self.chans):
+                l = c.geom_length/(2*mc.voxelsize[0])
+                x0, y0, z0 = c.centroid - l*c.principalAxis
+                x1, y1, z1 = c.centroid + l*c.principalAxis
+                
+                ax.plot([x0, x1], [y0, y1], [z0, z1],c = cols[i] , lw=3)
+                
+                x0, y0, z0 = c.centroid
+                ax.plot([x0], [y0], [z0], 'x', c = cols[i], lw=3)
+                
+        _drawaxes(ax)
+            
+
+            
+#        cols = [[1.,0.,0.], [0.,1.,0.], [0.,0.,1.]]
+#        for i, c in enumerate(self.chans):
+#            d = c.data
+#            verts, faces = measure.marching_cubes(d, isovalues[i])
+#            vertsc = verts + np.array(c.bbox[:3])/np.array(mc.voxelsize)
+#            tris =  vertsc[faces]
+#            #mesh = Poly3DCollection(vertsc[faces], facecolor=['r', 'g', 'b'][i], edgecolor=['r', 'g', 'b'][i], alpha=.5)
+#            #mesh = Poly3DCollection(tris, facecolor=self._shade(tris, cols[i]), edgecolor=['r', 'g', 'b'][i], alpha=.5)
+#            mesh = Poly3DCollection(tris, facecolor=self._shade(tris, cols[i]), edgecolor='none', alpha=.5)
+#            ax.add_collection3d(mesh)
+        
+        for tris, fcols in meshes:
+            mesh = Poly3DCollection(tris, facecolor=fcols, edgecolor='none', alpha=.5)
+            ax.add_collection3d(mesh)
+            
+
+        x0, y0, z0 = np.array(mc.bbox[:3])/np.array(mc.voxelsize)
+        x1, y1, z1 = np.array(mc.bbox[3:])/np.array(mc.voxelsize)
+
+        xc = (x0 + x1)/2
+        al = x1 - x0
+        yc = (y0 + y1)/2
+        al = max(al, y1 - y0)        
+        zc = (z0 + z1)/2
+        al = max(al, z1 - z0)
+        
+        al2 = al/2.0
+        
+        ax.set_xlim(xc - al2, xc + al2)
+        ax.set_ylim(yc - al2, yc + al2)
+        ax.set_zlim(zc - al2, zc + al2)        
+        #print c.centroid, vertsc
+        #ax.axis('equal')
+        #ax.set_axis_off()
+        
+        #ax = fig.add_subplot(1, 2, 2, projection='3d')
+        ax = Axes3D(fig, [.5, 0, .5,1], azim = 60, elev=30, axisbg='none')
+#        for m in meshes:
+#            ax.add_collection3d(m)
+        _drawaxes(ax)
+        
+        for tris, fcols in meshes:
+            mesh = Poly3DCollection(tris, facecolor=fcols, edgecolor='none', alpha=.5)
+            ax.add_collection3d(mesh)
+            
+        ax.set_xlim(xc - al2, xc + al2)
+        ax.set_ylim(yc - al2, yc + al2)
+        ax.set_zlim(zc - al2, zc + al2)
+        
+        #ax.view_init(-45, 45)
+        
+        out = BytesIO()
+        canvas.print_png(out, dpi=100, facecolor='none')
+        s = out.getvalue()
+        out.close()
+        
+        self._3diso = s
+        return s
+
+class Measurements(wx.Panel, Plugin):
+    def __init__(self, dsviewer):
+        wx.Panel.__init__(self, dsviewer)
+        Plugin.__init__(self, dsviewer)
+        
+        dsviewer.view.add_overlay(self.DrawOverlays, 'Blob measurements')
+        
+        vsizer=wx.BoxSizer(wx.VERTICAL)
+        
+        hsizer = wx.BoxSizer(wx.HORIZONTAL)
+        hsizer.Add(wx.StaticText(self, -1, 'Reference Channel:'), 0, wx.ALL|wx.ALIGN_CENTER_VERTICAL, 2)
+        
+        self.chMaster=wx.Choice(self, -1, choices=self.image.names)
+        self.chMaster.SetSelection(0)
+        hsizer.Add(self.chMaster, 1, wx.ALL|wx.ALIGN_CENTER_VERTICAL, 2)
+        
+        vsizer.Add(hsizer, 0, wx.EXPAND, 0)
+        
+        hsizer = wx.BoxSizer(wx.HORIZONTAL)
+        hsizer.Add(wx.StaticText(self, -1, 'Orient by:'), 0, wx.ALL|wx.ALIGN_CENTER_VERTICAL, 2)
+        
+        self.chOrient=wx.Choice(self, -1, choices=self.image.names)
+        self.chOrient.SetSelection(1)
+        hsizer.Add(self.chOrient, 1, wx.ALL|wx.ALIGN_CENTER_VERTICAL, 2)
+        
+        self.chDirection=wx.Choice(self, -1, choices=['-ve from reference', '+ve from reference'], size=(30, -1))
+        self.chDirection.SetSelection(0)
+        hsizer.Add(self.chDirection, 1, wx.ALL|wx.ALIGN_CENTER_VERTICAL, 2)
+        
+        vsizer.Add(hsizer, 0, wx.EXPAND, 0)
+        
+        sbsizer = wx.StaticBoxSizer(wx.StaticBox(self, -1, 'Minimum Amplitudes:'), wx.VERTICAL)
+        nChans = len(self.image.names)        
+        
+        gsizer = wx.FlexGridSizer(nChans, 2, 2, 5)
+        
+        self.aTexts = []
+        for chan in range(nChans):
+            gsizer.Add(wx.StaticText(self, -1, self.image.names[chan]), 1, wx.ALIGN_CENTER_VERTICAL)
+            at = wx.TextCtrl(self, -1, '0')
+            self.aTexts.append(at)
+            gsizer.Add(at, 2, wx.ALIGN_CENTER_VERTICAL)
+            
+        sbsizer.Add(gsizer, 0, wx.ALL|wx.EXPAND, 2)
+        
+        vsizer.Add(sbsizer, 0, wx.EXPAND, 2)
+        
+        bCalculate = wx.Button(self, -1, 'Measure')
+        bCalculate.Bind(wx.EVT_BUTTON, self.OnCalculate)
+        vsizer.Add(bCalculate, 0, wx.EXPAND|wx.ALL, 2)
+        
+        bHide = wx.Button(self, -1, 'Hide Object')
+        bHide.Bind(wx.EVT_BUTTON, self.OnHideObject)
+        vsizer.Add(bHide, 0, wx.EXPAND|wx.ALL, 2)
+        
+        bView = wx.Button(self, -1, 'View Results')
+        bView.Bind(wx.EVT_BUTTON, self.OnView)
+        vsizer.Add(bView, 0, wx.EXPAND|wx.ALL, 2)
+        
+        bXls = wx.Button(self, -1, 'Export to xls')
+        bXls.Bind(wx.EVT_BUTTON, self.OnExportXls)
+        vsizer.Add(bXls, 0, wx.EXPAND|wx.ALL, 2)
+        
+        self.SetSizerAndFit(vsizer)
+        
+        self.objects = []
+        self.ID = getNewID()
+        
+        self.StartServing()
+
+    def StartServing(self):
+        from PYME.DSView import htmlServe #ensure that our local cherrypy server is running
+        import cherrypy
+        
+        htmlServe.StartServing()
+        cherrypy.tree.mount(self, '/measure/%d' % self.ID)
+                    
+        
+#    def StartServing(self):
+#        try: 
+#            import threading
+#            self.serveThread = threading.Thread(target=self._serve)
+#            self.serveThread.start()
+#        except ImportError:
+#            pass
+#            
+#    def _serve(self):
+#        import cherrypy
+#        cherrypy.quickstart(self, '/measure/%d' % self.ID)
+        
+    def OnCalculate(self, event):
+        self.RetrieveObjects(self.chMaster.GetSelection(), self.chOrient.GetSelection(), 2*self.chDirection.GetSelection() - 1)
+        
+        cts = [float(at.GetValue()) for at in self.aTexts]
+        
+        for obj in self.objects:
+            for chan, ct in zip(obj.chans, cts):
+                if chan.sum < ct:
+                    obj.shown=False
+                    
+        self.dsviewer.do.OnChange()
+        
+    def OnHideObject(self, event):
+        do = self.dsviewer.do
+        ind = self.image.labels[do.xp, do.yp, do.zp]
+        
+        if ind:
+            self.objects[ind - 1].shown = False
+            do.OnChange()
+    
+    def OnView(self, event):
+        import webbrowser
+        from PYME.DSView import htmlServe
+        
+        webbrowser.open('%smeasure/%d' % (htmlServe.getURL(), self.ID))
+        
+    def OnSaveObjects(self, event):
+        #import cPickle
+        import zipfile
+        
+        zf = zipfile()
+        
+    def OnExportXls(self, event):
+        import os
+        filename = wx.FileSelector('Save to xls', default_filename=os.path.splitext(self.image.filename)[0] + '_blobMeasure.xls', wildcard='*.xls')
+        self.toXls(filename)
+            
+    
+    def GetRegion(self, index, objects = None):
+        if not objects:
+            objects = ndimage.find_objects(self.image.labels)
+            
+        o = objects[index]
+        
+        mask = self.image.labels[o] == (index+1)
+        
+        slx, sly, slz = o
+        
+        X, Y, Z = np.ogrid[slx, sly, slz]
+        vs = self.image.voxelsize
+        
+        #return [DataBlock(np.maximum(self.image.data[slx, sly, slz, j] - self.image.data[slx, sly, slz, j].min(), 0)*mask , X, Y, Z, vs) for j in range(self.image.data.shape[3])]
+        return [DataBlock(np.maximum(self.image.data[slx, sly, slz, j], 0)*mask , X, Y, Z, vs) for j in range(self.image.data.shape[3])]
+        
+    def RetrieveObjects(self, masterChan=0, orientChan=1, orient_dir=-1):
+        objs = ndimage.find_objects(self.image.labels)
+        
+        self.objects = [BlobObject(self.GetRegion(i, objs), masterChan, orientChan, orient_dir) for i in range(self.image.labels.max())]
+        
+    def DrawOverlays(self, view, dc):
+        import wx
+        col = wx.TheColourDatabase.FindColour('CYAN')
+        dc.SetPen(wx.Pen(col,2))
+        dc.SetBrush(wx.TRANSPARENT_BRUSH)
+        
+        for obj in self.objects:
+            obj.drawOverlay(view, dc)
+            
+        dc.SetPen(wx.NullPen)
+        dc.SetBrush(wx.NullBrush)
+        
+    def index(self, templateName='measureView3.html'):
+        from jinja2 import Environment, PackageLoader
+        env = Environment(loader=PackageLoader('PYME.DSView.modules', 'templates'))
+        
+        #if self.dsviewer.image.data.shape[2] > 1:
+            #3D image
+       #     templateName=
+        
+        template = env.get_template(templateName)
+        
+        return template.render(objects=self.objects)
+    index.exposed = True
+    
+    def _xlData(self, book, graphName):
+        gs = book.add_sheet(graphName) 
+        
+        #r = 3
+        
+        objects = [obj for obj in self.objects if obj.shown]
+        
+        #write the data first
+        c0 = 2
+        
+        for obj in objects:
+            xv, yv = getattr(obj, graphName)()
+            
+            
+            #for x_v in xv:
+            #    gs.write(r, c, x_v)
+            #    c += 1
+            
+            #add a gap
+            #c += 1
+            
+            c = c0
+            
+            
+            for yvc in yv:
+                gs.write(2, c, 'Obj %d' % self.objects.index(obj), self._boldStyle)
+                r = 3
+                for y_v in yvc:
+                    gs.write(r, c, y_v)
+                    r += 1
+                c += (len(objects) + 1)
+            
+            c0 += 1
+            #r+=1
+            
+        c = 0
+        r = 3
+        for x_v in xv:
+            gs.write(r, c, x_v)
+            r += 1
+            #    c += 1
+            
+                    
+        #now go back and write the header
+        #gs.merge(0, 0, 0, len(xv))
+        gs.write(0, 0, 'Bin edges', self._boldStyle)
+        for i, n in enumerate(self.image.names):
+            gs.merge( 0, 0,  2+ i*(len(objects)+1),  1 + (i+1)*(len(objects)+1))
+            gs.write(0, 2+ i*(len(objects)+1) , n, self._boldStyle)    
+        
+    def toXls(self, filename):
+        import xlwt
+        
+        book = xlwt.Workbook()
+        
+        gs = book.add_sheet('General')
+        
+        #set up bold style
+        self._boldStyle = xlwt.easyxf('font: bold 1')
+        
+        paramsToWrite=['Centroid x','Centroid y', 'Length', 'Width', 'Sum']
+        nParams = len(paramsToWrite)
+        
+        #Header
+        gs.write(1,0,'Image #', self._boldStyle)
+        for i, n in enumerate(self.image.names):
+            gs.merge( 0, 0, 1+ i*nParams, (i+1)*nParams,)
+            gs.write(0, 1+ i*nParams, n, self._boldStyle)
+            for j, p in enumerate(paramsToWrite):
+                gs.write(1, 1 + i*nParams + j, p, self._boldStyle)
+        
+        r = 2
+        for obj in self.objects:
+            if obj.shown:
+                gs.write(r, 0, self.objects.index(obj))
+                for i, ch in enumerate(obj.chans):
+                    gs.write(r, 1 + i*nParams + 0, ch.centroid[0])
+                    gs.write(r, 1 + i*nParams + 1, ch.centroid[1])
+                    gs.write(r, 1 + i*nParams + 2, ch.mad_0*2.35/0.8)
+                    gs.write(r, 1 + i*nParams + 3, ch.mad_1*2.35/0.8)
+                    gs.write(r, 1 + i*nParams + 4, float(ch.sum))
+                
+                r += 1
+         
+        r+=2
+        gs.write(r, 0, 'Reference Channel:', self._boldStyle)
+        gs.write(r, 1, self.image.names[self.objects[0].masterChan])
+                
+        #now do the graph data
+        graphs = ['shortAxisDist', 'longAxisDistN', 'angularDist', 'radialDistN', 'longAxisOrthDist']
+        
+        for gn in graphs:
+            self._xlData(book, gn)
+                
+        book.save(filename)
+        
+    
+    def images(self, num):
+        return self.objects[int(num)].getImage(self.dsviewer.do.Gains)
+    images.exposed = True
+    
+    def graphs(self, num, graphName):
+        return self.objects[int(num)].getGraph(graphName)
+    graphs.exposed = True
+    
+    def schemes(self, num):
+        return self.objects[int(num)].getSchematic()
+    schemes.exposed = True
+    
+    def schemes3D(self, num):
+        return self.objects[int(num)].getSchematic3D()
+    schemes3D.exposed = True
+    
+    def isosurface3D(self, num):
+        return self.objects[int(num)].get3DIsosurf(0.5/np.array(self.dsviewer.do.Gains))
+    isosurface3D.exposed = True
+    
+    def hide(self, num):
+        self.objects[int(num)].shown = False
+        self.dsviewer.do.OnChange()
+        return ''
+    hide.exposed = True
+    
+        
+        
+def Plug(dsviewer):
+    measure = Measurements(dsviewer)
+    
+    measure.SetSize(measure.GetBestSize())
+    pinfo2 = aui.AuiPaneInfo().Name("measurePanel").Left().Caption('Blob Characterisation').CloseButton(False).MinimizeButton(True).MinimizeMode(aui.AUI_MINIMIZE_CAPT_SMART|aui.AUI_MINIMIZE_POS_RIGHT)#.CaptionVisible(False)
+    dsviewer._mgr.AddPane(measure, pinfo2)
+    dsviewer._mgr.Update()
+    
+    return measure