Changes between Version 35 and Version 36 of barSoftwareErrorCorrection
- Timestamp:
- 10/26/11 09:40:54 (13 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
barSoftwareErrorCorrection
v35 v36 7 7 In our work-flow we introduced a stage of ''contour slides'' - vector graphic in SVG format (for detailed description see Majka et al 2011) that simplifies necessary corrections, manual preparation of structures' outlines from histology results and further improvements of slide details. 8 8 9 '''FIGURE 1.''' [[BR]]9 [=#figure1]'''FIGURE 1.''' [[BR]] 10 10 '''Left''': Coronal slice of a rat brain around 3 mm behind Bregma stained for Cytochrome Oxidase. Histological verification of the electrodes' location after chronic local field potentials (LFP) recording from the thalamic somatosensory nuclei. '''Middle''': Initial stage of manual preparation of a ''countour slide''. Main structures outlines drawn over the histology slice (line thickness increased for better visibility) and structure labels placed over well recognized regions. '''Right''': Histological slice visible through translucent final CAF slide. 11 11 [[Image(006__final_comparison.png)]] … … 45 45 A problem we encountered during tracing was the lack of labels 46 46 attributed to some areas. This can happen when "obvious" names are 47 omitted in commercial atlases; when user does not recognize the area in histological slice (see example in figure 5), but also when additional area is erroneously defined by accidental line crossing (see figure 2below).47 omitted in commercial atlases; when user does not recognize the area in histological slice (see example in [#figure5 figure 5]), but also when additional area is erroneously defined by accidental line crossing (see [#figure2 figure 2] below). 48 48 49 49 3dBAR vector parser automatically detects such unlabeled regions by comparing the sum of traced paths with the total area of the ''Brain'' structure. The parser locates all … … 53 53 the regular structures. 54 54 55 '''FIGURE 2.''' [[BR]]56 '''Left''': Red blocks mark regions where irregular neighboring lines touch each other creating small closed areas. '''Middle''': Small areas were detected by the parser, traced and marked as Unlabelled. '''Right''': Adjusting the threshold number of pixels that should be trace sby parser excluded incorrectly recognized tiny structures.55 [=#figure2]'''FIGURE 2.''' [[BR]] 56 '''Left''': Red blocks mark regions where irregular neighboring lines touch each other creating small closed areas. '''Middle''': Small areas were detected by the parser, traced and marked as Unlabelled. '''Right''': Adjusting the threshold number of pixels that should be traced by parser excluded incorrectly recognized tiny structures. 57 57 [[Image(001__unlabelled_annotated.gif)]] 58 58 … … 69 69 about ambiguous delineation and should review the slide in question 70 70 and correct it.[[BR]] 71 In the CAF slide on the right of Figure 2above we can see two labels (ic, Rt) located within one region. One of them was changed by parser to spot label (.ic). This particular example resulted from improper drawing and a gap existing in the border between internal capsule (ic) and reticular nucleus of the thalamus (Rt). The parser treated areas connected by the gap as one structure (''leaking of the structure'', see below).71 In the CAF slide on the right of [#figure2 figure 2] above we can see two labels (ic, Rt) located within one region. One of them was changed by parser to spot label (.ic). This particular example resulted from improper drawing and a gap existing in the border between internal capsule (ic) and reticular nucleus of the thalamus (Rt). The parser treated areas connected by the gap as one structure (''leaking of the structure'', see below). 72 72 73 73 === Incorrect label placement === … … 79 79 Tracing of a region corresponding to the incorrectly placed label is skipped but in general tracing is not interrupted as the whole procedure works non-interactively. 80 80 81 '''FIGURE 3.''' [[BR]]81 [=#figure3]'''FIGURE 3.''' [[BR]] 82 82 [[Image(003__invalid_labels_annotated.gif)]] 83 83 ---- … … 92 92 structure being traced overtakes the space of its neighbour through 93 93 the broken contour which we call ''leaking of the structure'' (see 94 inserts in Figure 4 below).94 inserts in [#figure4 figure 4]). 95 95 96 96 It is handled by a heuristic gap filling algorithm ([wiki:barSoftwareGapFillDetails detailed description]). The main idea behind this algorithm is to expand the contours by applying a [http://en.wikipedia.org/wiki/Mathematical_morphology dilation] filter … … 101 101 tremendous as compared with manual cleaning. 102 102 103 '''FIGURE 4.'''[[BR]]103 [=#figure4]'''FIGURE 4.'''[[BR]] 104 104 [[Image(002__gapfill_annotated.gif)]] 105 105 ---- … … 112 112 Unrecognized areas can be attributed to the structures from the upper hierarchy level. 113 113 114 '''FIGURE 5.'''[[BR]]114 [=#figure5]'''FIGURE 5.'''[[BR]] 115 115 '''Left''': New lines and appropriate labels (blinking) were added to the ''contour slide'' 1) to show primary somatosensory representation of big facial whiskers ("barrel field", S1BF) in the cerebral cortex (Cx); and 2) to divide ventrobasal complex (VB) into ventral posteromedial (VPM) and ventral posterolateral (VPL) nuclei. Unrecognized thalamic nuclei, which were not labeled with their specific names, are labeled with "Th" so they all will be included in the final reconstruction of the thalamus. '''Right''': Updated CAF slide. 116 116 [[Image(0041_new_edges_animated.gif)]] … … 120 120 In our example we include in the slide an outline of the electrode. Spaces occupied by lesions or injected dyes can be also added to atlas elements. 121 121 122 '''FIGURE 6.'''[[BR]]122 [=#figure6]'''FIGURE 6.'''[[BR]] 123 123 [[Image(007__final_electrode.png)]] 124 124 ----
