| 1. |
- Wang, Chunliang, et al.
(författare)
-
Automatic heart and vessel segmentation using random forests and a local phase guided level set method
- 2017
-
Ingår i: Reconstruction, Segmentation, and Analysis Of Medical Images. - Cham : Springer Verlag. - 9783319522791 ; , s. 159-164
-
Konferensbidrag (refereegranskat)abstract
- In this report, a novel automatic heart and vessel segmentation method is proposed. The heart segmentation pipeline consists of three major steps: heart localization using landmark detection, heart isolation using statistical shape model and myocardium segmentation using learning based voxel classification and local phase analysis. In our preliminary test, the proposed method achieved encouraging results.
|
|
| 2. |
- Astaraki, Mehdi, PhD Student, 1984-, et al.
(författare)
-
Multimodal brain tumor segmentation with normal appearance autoencoder
- 2019
-
Ingår i: International MICCAI Brainlesion Workshop. - Cham : Springer Nature. ; , s. 316-323
-
Konferensbidrag (refereegranskat)abstract
- We propose a hybrid segmentation pipeline based on the autoencoders’ capability of anomaly detection. To this end, we, first, introduce a new augmentation technique to generate synthetic paired images. Gaining advantage from the paired images, we propose a Normal Appearance Autoencoder (NAA) that is able to remove tumors and thus reconstruct realistic-looking, tumor-free images. After estimating the regions where the abnormalities potentially exist, a segmentation network is guided toward the candidate region. We tested the proposed pipeline on the BraTS 2019 database. The preliminary results indicate that the proposed model improved the segmentation accuracy of brain tumor subregions compared to the U-Net model.
|
|
| 3. |
- Astaraki, Mehdi, PhD Student, 1984-, et al.
(författare)
-
Normal appearance autoencoder for lung cancer detection and segmentation
- 2019
-
Ingår i: International Conference on Medical Image Computing and Computer-Assisted Intervention. - Cham : Springer Nature. ; , s. 249-256
-
Konferensbidrag (refereegranskat)abstract
- One of the major differences between medical doctor training and machine learning is that doctors are trained to recognize normal/healthy anatomy first. Knowing the healthy appearance of anatomy structures helps doctors to make better judgement when some abnormality shows up in an image. In this study, we propose a normal appearance autoencoder (NAA), that removes abnormalities from a diseased image. This autoencoder is semi-automatically trained using another partial convolutional in-paint network that is trained using healthy subjects only. The output of the autoencoder is then fed to a segmentation net in addition to the original input image, i.e. the latter gets both the diseased image and a simulated healthy image where the lesion is artificially removed. By getting access to knowledge of how the abnormal region is supposed to look, we hypothesized that the segmentation network could perform better than just being shown the original slice. We tested the proposed network on the LIDC-IDRI dataset for lung cancer detection and segmentation. The preliminary results show the NAA approach improved segmentation accuracy substantially in comparison with the conventional U-Net architecture.
|
|
| 4. |
- Bendazzoli, Simone, et al.
(författare)
-
Automatic rat brain segmentation from MRI using statistical shape models and random forest
- 2019
-
Ingår i: MEDICAL IMAGING 2019. - : SPIE-INT SOC OPTICAL ENGINEERING. - 9781510625464 - 9781510625457
-
Konferensbidrag (refereegranskat)abstract
- In MRI neuroimaging, the shimming procedure is used before image acquisition to correct for inhomogeneity of the static magnetic field within the brain. To correctly adjust the field, the brain's location and edges must first be identified from quickly-acquired low resolution data. This process is currently carried out manually by an operator, which can be time-consuming and not always accurate. In this work, we implement a quick and automatic technique for brain segmentation to be potentially used during the shimming. Our method is based on two main steps. First, a random forest classifier is used to get a preliminary segmentation from an input MRI image. Subsequently, a statistical shape model of the brain, which was previously generated from ground-truth segmentations, is fitted to the output of the classifier to obtain a model-based segmentation mask. In this way, a-priori knowledge on the brain's shape is included in the segmentation pipeline. The proposed methodology was tested on low resolution images of rat brains and further validated on rabbit brain images of higher resolution. Our results suggest that the present method is promising for the desired purpose in terms of time efficiency, segmentation accuracy and repeatability. Moreover, the use of shape modeling was shown to be particularly useful when handling low-resolution data, which could lead to erroneous classifications when using only machine learning-based methods.
|
|
| 5. |
- Brusini, Irene, et al.
(författare)
-
Fully automatic estimation of the waist of the nerve fiber layer at the optic nerve head angularly resolved
- 2021
-
Ingår i: Progress in Biomedical Optics and Imaging - Proceedings of SPIE. - : SPIE-Intl Soc Optical Eng. ; , s. 1D1-1D8
-
Konferensbidrag (refereegranskat)abstract
- The present project aims at developing a fully automatic software for estimation of the waist of the nerve fiber layer in the Optic Nerve Head (ONH) angularly resolved in the frontal plane as a tool for morphometric monitoring of glaucoma. The waist of the nerve fiber layer is here defined as Pigment epithelium central limit –Inner limit of the retina – Minimal Distance, (PIMD). 3D representations of the ONH were collected with high resolution OCT in young not glaucomatous eyes and glaucomatous eyes. An improved tool for manual annotation was developed in Python. This tool was found user friendly and to provide sufficiently precise manual annotation. PIMD was automatically estimated with a software consisting of one AI model for detection of the inner limit of the retina and another AI model for localization of the Optic nerve head Pigment epithelium Central limit (OPCL). In the current project, the AI model for OPCL localization was retrained with new data manually annotated with the improved tool for manual annotation both in not glaucomatous eyes and in glaucomatous eyes. Finally, automatic annotations were compared to 3 annotations made by 3 independent annotators in an independent subset of both the not glaucomatous and the glaucomatous eyes. It was found that the fully automatic estimation of PIMD-angle overlapped the 3 manual annotators with small variation among the manual annotators. Considering interobserver variation, the improved tool for manual annotation provided less variation than our original annotation tool in not glaucomatous eyes suggesting that variation in glaucomatous eyes is due to variable pathological anatomy, difficult to annotate without error. The small relative variation in relation to the substantial overall loss of PIMD in the glaucomatous eyes compared to the not glaucomatous eyes suggests that our software for fully automatic estimation of PIMD-angle can now be implemented clinically for monitoring of glaucoma progression.
|
|
| 6. |
|
|
| 7. |
- Carrizo, Garrizo, et al.
(författare)
-
Fully automatic estimation of the angular distribution of the waist of the nerve fiber layer in the optic nerve head
- 2020
-
Ingår i: Ophthalmic Technologies XXX. - : SPIE-Intl Soc Optical Eng.
-
Konferensbidrag (refereegranskat)abstract
- In this paper, an automatic strategy for measuring the thickness of the nerve fiber layer around the optic nerve head is proposed. The strategy presented uses two independent 2D U-nets that each perform a segmentation task. One network learns to segment the vitreous body in standard Cartesian image domain and the second learns to segment a disc around a point of interest in a polar image domain. The output from the neural networks are then combined to find the thickness of the waist of the nerve fiber layer as a function of the angle around the center of the optic nerve head in the frontal plane. The two networks are trained with a combined data set that has been captured on two separate OCT systems (spectral domain Topcon OCT 2000 and swept source Topcon OCT Triton) which have been annotated with a semi-automatic algorithm by up to 3 annotators. Initial results show that the automatic algorithm produces results that are comparable to the results from the semi-automatic algorithm used for reference, in a fraction of the time, independent of the annotator. The automatic algorithm has the potential to replace the semi-automatic algorithm and opens the possibility for clinical routine estimation of the nerve fiber layer. This would in turn allow the detection of loss of nerve fiber layer earlier than before which is anticipated to be important for detection of glaucoma.
|
|
| 8. |
- Chowdhury, Manish, et al.
(författare)
-
Segmentation of Cortical Bone using Fast Level Sets
- 2017
-
Ingår i: MEDICAL IMAGING 2017. - : SPIE - International Society for Optical Engineering. - 9781510607118
-
Konferensbidrag (refereegranskat)abstract
- Cortical bone plays a big role in the mechanical competence of bone. The analysis of cortical bone requires accurate segmentation methods. Level set methods are usually in the state-of-the-art for segmenting medical images. However, traditional implementations of this method are computationally expensive. This drawback was recently tackled through the so-called coherent propagation extension of the classical algorithm which has decreased computation times dramatically. In this study, we assess the potential of this technique for segmenting cortical bone in interactive time in 3D images acquired through High Resolution peripheral Quantitative Computed Tomography (HR-pQCT). The obtained segmentations are used to estimate cortical thickness and cortical porosity of the investigated images. Cortical thickness and Cortical porosity is computed using sphere fitting and mathematical morphological operations respectively. Qualitative comparison between the segmentations of our proposed algorithm and a previously published approach on six images volumes reveals superior smoothness properties of the level set approach. While the proposed method yields similar results to previous approaches in regions where the boundary between trabecular and cortical bone is well defined, it yields more stable segmentations in challenging regions. This results in more stable estimation of parameters of cortical bone. The proposed technique takes few seconds to compute, which makes it suitable for clinical settings.
|
|
| 9. |
- Kisonaite, Konstancija, et al.
(författare)
-
Estimation of the cross-sectional surface area of the waist of the nerve fiber layer at the optic nerve head
- 2022
-
Ingår i: Progress in Biomedical Optics and Imaging. - : SPIE-Intl Soc Optical Eng.
-
Konferensbidrag (refereegranskat)abstract
- Glaucoma is a global disease that leads to blindness due to pathological loss of retinal ganglion cell axons in the optic nerve head (ONH). The presented project aims at improving a computational algorithm for estimating the thickness and surface area of the waist of the nerve fiber layer in the ONH. Our currently developed deep learning AI algorithm meets the need for a morphometric parameter that detects glaucomatous change earlier than current clinical follow-up methods. In 3D OCT image volumes, two different AI algorithms identify the Optic nerve head Pigment epithelium Central Limit (OPCL) and the Inner limit of the Retina Closest Point (IRCP) in a 3D grid. Our computational algorithm includes the undulating surface area of the waist of the ONH, as well as waist thickness. In 16 eyes of 16 non-glaucomatous subjects aged [20;30] years, the mean difference in minimal thickness of the waist of the nerve fiber layer between our previous and the current post-processing strategies was estimated as CIμ(0.95) 0 ±1 μm (D.f. 15). The mean surface area of the waist of the nerve fiber layer in the optic nerve head was 1.97 ± 0.19 mm2. Our computational algorithm results in slightly higher values for surface areas compared to published work, but as expected, this may be due to surface undulations of the waist being considered. Estimates of the thickness of the waist of the ONH yields estimates of the same order as our previous computational algorithm.
|
|
| 10. |
- Lidayová, Kristína, et al.
(författare)
-
Coverage segmentation of 3D thin structures
- 2015
-
Ingår i: Image Processing Theory, Tools and Applications (IPTA), 2015 International Conference on. - Piscataway, NJ : IEEE conference proceedings. - 9781479986361 ; , s. 23-28
-
Konferensbidrag (refereegranskat)abstract
- We present a coverage segmentation method for extracting thin structures in three-dimensional images. The proposed method is an improved extension of our coverage segmentation method for 2D thin structures. We suggest implementation that enables low memory consumption and processing time, and by that applicability of the method on real CTA data. The method needs a reliable crisp segmentation as an input and uses information from linear unmixing and the crisp segmentation to create a high-resolution crisp reconstruction of the object, which can then be used as a final result, or down-sampled to a coverage segmentation at the starting image resolution. Performed quantitative and qualitative analysis confirm excellent performance of the proposed method, both on synthetic and on real data, in particular in terms of robustness to noise.
|
|
