Limits...
2D Fast Vessel Visualization Using a Vessel Wall Mask Guiding Fine Vessel Detection.

Raptis S, Koutsouris D - Int J Biomed Imaging (2010)

Bottom Line: Sparse pixels are effectively eliminated by applying a limited range Hough Transform (HT) or region growing.Major benefits are limiting the range of parameters, reducing the search-space for post-convolution to only masked regions, representing almost 2% of the 2D volume, good speed versus accuracy/time trade-off.Results show the potentials of our approach in terms of time for detection ROC analysis and accuracy of vessel pixel (VP) detection.

View Article: PubMed Central - PubMed

Affiliation: Biomedical Engineering Laboratory, School of Electrical & Computer Engineering, National Technical University of Athens, 9 Iroon Polytechniou Str., H/Y Building-Zografou Campus, 15773 Athens, Greece.

ABSTRACT
The paper addresses the fine retinal-vessel's detection issue that is faced in diagnostic applications and aims at assisting in better recognizing fine vessel anomalies in 2D. Our innovation relies in separating key visual features vessels exhibit in order to make the diagnosis of eventual retinopathologies easier to detect. This allows focusing on vessel segments which present fine changes detectable at different sampling scales. We advocate that these changes can be addressed as subsequent stages of the same vessel detection procedure. We first carry out an initial estimate of the basic vessel-wall's network, define the main wall-body, and then try to approach the ridges and branches of the vasculature's using fine detection. Fine vessel screening looks into local structural inconsistencies in vessels properties, into noise, or into not expected intensity variations observed inside pre-known vessel-body areas. The vessels are first modelled sufficiently but not precisely by their walls with a tubular model-structure that is the result of an initial segmentation. This provides a chart of likely Vessel Wall Pixels (VWPs) yielding a form of a likelihood vessel map mainly based on gradient filter's intensity and spatial arrangement parameters (e.g., linear consistency). Specific vessel parameters (centerline, width, location, fall-away rate, main orientation) are post-computed by convolving the image with a set of pre-tuned spatial filters called Matched Filters (MFs). These are easily computed as Gaussian-like 2D forms that use a limited range sub-optimal parameters adjusted to the dominant vessel characteristics obtained by Spatial Grey Level Difference statistics limiting the range of search into vessel widths of 16, 32, and 64 pixels. Sparse pixels are effectively eliminated by applying a limited range Hough Transform (HT) or region growing. Major benefits are limiting the range of parameters, reducing the search-space for post-convolution to only masked regions, representing almost 2% of the 2D volume, good speed versus accuracy/time trade-off. Results show the potentials of our approach in terms of time for detection ROC analysis and accuracy of vessel pixel (VP) detection.

No MeSH data available.


Related in: MedlinePlus

(a) Original fundus image for averagely  traced major and peripheral network along with drusen. (b) Original fundus image ellipse corrected. (c) Binary vessel wall map. (d) Binary fundus image ellipse and drusen corrected. (e) Retinal vasculature with coarse and fine network based on VWP. (f) Retinal vasculature without VWP map using MIP. (g) Manually labelled ground truth vessel vasculature. (h) Redundant vessel vasculature produced using an augmented VM and double-sided GS-MG. (i) 2nd vessel network: Successive maps taken using different features and different sensitivities (thresholds). (8) Gradient homogeneity VMs. (9) Paired VP map and dominant gradient VM. (10) Increased vessel width allowed, lower gradient threshold and interpolated WPs. (11) Experimentally proved VM.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2913788&req=5

fig3: (a) Original fundus image for averagely traced major and peripheral network along with drusen. (b) Original fundus image ellipse corrected. (c) Binary vessel wall map. (d) Binary fundus image ellipse and drusen corrected. (e) Retinal vasculature with coarse and fine network based on VWP. (f) Retinal vasculature without VWP map using MIP. (g) Manually labelled ground truth vessel vasculature. (h) Redundant vessel vasculature produced using an augmented VM and double-sided GS-MG. (i) 2nd vessel network: Successive maps taken using different features and different sensitivities (thresholds). (8) Gradient homogeneity VMs. (9) Paired VP map and dominant gradient VM. (10) Increased vessel width allowed, lower gradient threshold and interpolated WPs. (11) Experimentally proved VM.

Mentions: Some representative images are shown in (Figures 1–4)* where the asterisk “∗” indicates all subfigures or experiments referring to the same case. In (Figure 1)*, we show a not very clear background and a vessel network without any structural changes (drusen), with a very light background that does allow clearly seeing the peripheral vessel network details. In (Figure 3)* the vessel network has a more detailed peripheral vasculature and is sufficiently clear in that but still has some limited drusen, while in (Figure 2)* the background has a large amount of extended drusen. In (Figure 1)* and all cases, we show the original vessel network for all 3 vasculature types. Intermediate processing stages that are generated in the process are also shown in a figure series “∗” for all vasculature cases. These are produced by means of the wall detection mask and show the wall pixels, guiding the detection process. All stages are depicted with best candidate pixels. For visualization purposes wall pixels both give an idea of the basic network and also can alert on any evident missleading topologies like seriously broken segments or even abrupt cuts. Still, sparse pixels got separated using a finite, maximally 10-step region growing segmentation as in (Figure 5)* that lasted 10–15 seconds for the entire image. When a grown region recruited only a small number of pixels, typically (10≥15) then these pixels were set to the background class. The map methodology was examined in terms of time, false positives (FP's) and true positive (TP's) related ratios and all associated ROC analysis performers as will be explained later in this section. Pixel pairing, done with the use of a VWP map, removes significantly point characteristics very similar to vessel like point characteristics. The interior vessel-like network is thus limited in Figure 3(h) As it is shown in the case of (Figure 3)* the drussen have been algorithmically corrected. This was achieved by applying successive MFR threshold adaptation and local geometrical checking of the resulting pixels locations as described in the gradient coherence formula-criterion that is discussed in the Sections 2.2 and 4. We can then see the development of the results when spatial gradient coherence is applied as discussed where the spatial continuity of the gradient is imposed.


2D Fast Vessel Visualization Using a Vessel Wall Mask Guiding Fine Vessel Detection.

Raptis S, Koutsouris D - Int J Biomed Imaging (2010)

(a) Original fundus image for averagely  traced major and peripheral network along with drusen. (b) Original fundus image ellipse corrected. (c) Binary vessel wall map. (d) Binary fundus image ellipse and drusen corrected. (e) Retinal vasculature with coarse and fine network based on VWP. (f) Retinal vasculature without VWP map using MIP. (g) Manually labelled ground truth vessel vasculature. (h) Redundant vessel vasculature produced using an augmented VM and double-sided GS-MG. (i) 2nd vessel network: Successive maps taken using different features and different sensitivities (thresholds). (8) Gradient homogeneity VMs. (9) Paired VP map and dominant gradient VM. (10) Increased vessel width allowed, lower gradient threshold and interpolated WPs. (11) Experimentally proved VM.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2913788&req=5

fig3: (a) Original fundus image for averagely traced major and peripheral network along with drusen. (b) Original fundus image ellipse corrected. (c) Binary vessel wall map. (d) Binary fundus image ellipse and drusen corrected. (e) Retinal vasculature with coarse and fine network based on VWP. (f) Retinal vasculature without VWP map using MIP. (g) Manually labelled ground truth vessel vasculature. (h) Redundant vessel vasculature produced using an augmented VM and double-sided GS-MG. (i) 2nd vessel network: Successive maps taken using different features and different sensitivities (thresholds). (8) Gradient homogeneity VMs. (9) Paired VP map and dominant gradient VM. (10) Increased vessel width allowed, lower gradient threshold and interpolated WPs. (11) Experimentally proved VM.
Mentions: Some representative images are shown in (Figures 1–4)* where the asterisk “∗” indicates all subfigures or experiments referring to the same case. In (Figure 1)*, we show a not very clear background and a vessel network without any structural changes (drusen), with a very light background that does allow clearly seeing the peripheral vessel network details. In (Figure 3)* the vessel network has a more detailed peripheral vasculature and is sufficiently clear in that but still has some limited drusen, while in (Figure 2)* the background has a large amount of extended drusen. In (Figure 1)* and all cases, we show the original vessel network for all 3 vasculature types. Intermediate processing stages that are generated in the process are also shown in a figure series “∗” for all vasculature cases. These are produced by means of the wall detection mask and show the wall pixels, guiding the detection process. All stages are depicted with best candidate pixels. For visualization purposes wall pixels both give an idea of the basic network and also can alert on any evident missleading topologies like seriously broken segments or even abrupt cuts. Still, sparse pixels got separated using a finite, maximally 10-step region growing segmentation as in (Figure 5)* that lasted 10–15 seconds for the entire image. When a grown region recruited only a small number of pixels, typically (10≥15) then these pixels were set to the background class. The map methodology was examined in terms of time, false positives (FP's) and true positive (TP's) related ratios and all associated ROC analysis performers as will be explained later in this section. Pixel pairing, done with the use of a VWP map, removes significantly point characteristics very similar to vessel like point characteristics. The interior vessel-like network is thus limited in Figure 3(h) As it is shown in the case of (Figure 3)* the drussen have been algorithmically corrected. This was achieved by applying successive MFR threshold adaptation and local geometrical checking of the resulting pixels locations as described in the gradient coherence formula-criterion that is discussed in the Sections 2.2 and 4. We can then see the development of the results when spatial gradient coherence is applied as discussed where the spatial continuity of the gradient is imposed.

Bottom Line: Sparse pixels are effectively eliminated by applying a limited range Hough Transform (HT) or region growing.Major benefits are limiting the range of parameters, reducing the search-space for post-convolution to only masked regions, representing almost 2% of the 2D volume, good speed versus accuracy/time trade-off.Results show the potentials of our approach in terms of time for detection ROC analysis and accuracy of vessel pixel (VP) detection.

View Article: PubMed Central - PubMed

Affiliation: Biomedical Engineering Laboratory, School of Electrical & Computer Engineering, National Technical University of Athens, 9 Iroon Polytechniou Str., H/Y Building-Zografou Campus, 15773 Athens, Greece.

ABSTRACT
The paper addresses the fine retinal-vessel's detection issue that is faced in diagnostic applications and aims at assisting in better recognizing fine vessel anomalies in 2D. Our innovation relies in separating key visual features vessels exhibit in order to make the diagnosis of eventual retinopathologies easier to detect. This allows focusing on vessel segments which present fine changes detectable at different sampling scales. We advocate that these changes can be addressed as subsequent stages of the same vessel detection procedure. We first carry out an initial estimate of the basic vessel-wall's network, define the main wall-body, and then try to approach the ridges and branches of the vasculature's using fine detection. Fine vessel screening looks into local structural inconsistencies in vessels properties, into noise, or into not expected intensity variations observed inside pre-known vessel-body areas. The vessels are first modelled sufficiently but not precisely by their walls with a tubular model-structure that is the result of an initial segmentation. This provides a chart of likely Vessel Wall Pixels (VWPs) yielding a form of a likelihood vessel map mainly based on gradient filter's intensity and spatial arrangement parameters (e.g., linear consistency). Specific vessel parameters (centerline, width, location, fall-away rate, main orientation) are post-computed by convolving the image with a set of pre-tuned spatial filters called Matched Filters (MFs). These are easily computed as Gaussian-like 2D forms that use a limited range sub-optimal parameters adjusted to the dominant vessel characteristics obtained by Spatial Grey Level Difference statistics limiting the range of search into vessel widths of 16, 32, and 64 pixels. Sparse pixels are effectively eliminated by applying a limited range Hough Transform (HT) or region growing. Major benefits are limiting the range of parameters, reducing the search-space for post-convolution to only masked regions, representing almost 2% of the 2D volume, good speed versus accuracy/time trade-off. Results show the potentials of our approach in terms of time for detection ROC analysis and accuracy of vessel pixel (VP) detection.

No MeSH data available.


Related in: MedlinePlus