presentation project computer vision - teeth segmentation

26
Project for the Computer Vision course By Anne Everars

Upload: anne-everars

Post on 09-May-2015

286 views

Category:

Education


1 download

TRANSCRIPT

Page 1: Presentation project Computer Vision - Teeth segmentation

Project for the Computer Vision course By Anne Everars

Page 2: Presentation project Computer Vision - Teeth segmentation

� � Preprocessing

� Grayscaling � Region of interest (ROI) �  Separation of jaws � Adjust the image �  Separation of teeth

� Teeth segmentation � Water shedding � Mean shift filtering

� Teeth classification

Content

Page 3: Presentation project Computer Vision - Teeth segmentation

� Preprocessing

▶ Grayscaling ▶ Region of interest (ROI) ▶ Separation of jaws ▶ Adjust the image ▶ Separation of teeth

Page 4: Presentation project Computer Vision - Teeth segmentation

� �  Startpoint:

�  From RGB ➡ grayscale

Grayscaling

Page 5: Presentation project Computer Vision - Teeth segmentation

� � Image contains a lot of extra information (jaw, nose,

etc.) ➠ crop to a region of interest (ROI)

� Characteristics of the radiographs: � Variation in scale and position of head and teeth is

limited � Head is always centered horizontally

� Define the ROI based on the mean of the Gaussian distribution of the ROI of each image

Region of interest (ROI)

Page 6: Presentation project Computer Vision - Teeth segmentation

� Region of interest (ROI)

Page 7: Presentation project Computer Vision - Teeth segmentation

� � Teeth have a higher grey level intensity than jaws

and other (soft) tissue, because of their higher tissue density ➠ gap between jaws forms a valley in the y-axis projection histogram

� How? � Determine a set of points that have minimal intensity � Use interpolation to estimate the gap valley � Determine a split line (parallel to the x-axis)

Separation of jaws

Page 8: Presentation project Computer Vision - Teeth segmentation

� Separation of jaws

Page 9: Presentation project Computer Vision - Teeth segmentation

� Separation of jaws

Page 10: Presentation project Computer Vision - Teeth segmentation

� � Gaussian blurring ➠ more homogeneous

� Adaptive thresholding � Opening and closing morphological operations ➠ reduce noise

Adjust the image

Page 11: Presentation project Computer Vision - Teeth segmentation

� Gaussian blurring

Page 12: Presentation project Computer Vision - Teeth segmentation

� Adaptive thresholding

Page 13: Presentation project Computer Vision - Teeth segmentation

� Opening and closing

Page 14: Presentation project Computer Vision - Teeth segmentation

� = Search for maximum intensity in the y-direction

Separation of teeth

Page 15: Presentation project Computer Vision - Teeth segmentation

� Teeth segmentation

▶ Water shedding ▶ Mean shift filtering

Page 16: Presentation project Computer Vision - Teeth segmentation

� � Segment image based on similar intensity

� No proper segmentation : toothwas not sufficiently delineated ➠exterior is also flooded

Water shedding

Page 17: Presentation project Computer Vision - Teeth segmentation

� = (Partial) solution to the previous problem:

� Remove part of the noise (upper part of the image)

Mean shift filtering

Page 18: Presentation project Computer Vision - Teeth segmentation

� = (Partial) solution to the previous problem:

� Remove part of the noise (upper part of the image) � Apply a Gaussian blurring (again)

Mean shift filtering

Page 19: Presentation project Computer Vision - Teeth segmentation

� = (Partial) solution to the previous problem:

� Remove part of the noise (upper part of the image) � Apply a Gaussian blurring (again) � Apply mean shift filtering to smoothen the image

Mean shift filtering

Page 20: Presentation project Computer Vision - Teeth segmentation

� = (Partial) solution to the previous problem:

� Remove part of the noise (upper part of the image) � Apply a Gaussian blurring (again) � Apply mean shift filtering to smoothen the image � Reapply the water shedding algorithm (and adjust the

image)

Mean shift filtering

Page 21: Presentation project Computer Vision - Teeth segmentation

� Teeth classification

▶ Hamming distance ▶ Eigenfaces ▶ Principal Component Analysis (PCA)

Page 22: Presentation project Computer Vision - Teeth segmentation

� � Both methods do not work really well � The delineation is not good enough to perform a

good classification � Not implemented

� Possible methods to consider: � Hamming distance � Eigenfaces � Principal Component Analysis (PCA)

Results from segmentation

Page 23: Presentation project Computer Vision - Teeth segmentation

� � Use segmented image as a mask to compare with the

retrieved segmentation = compare

� Determine common scale

�  E.g. Smallest box arround both teeth � Determine number of not-matching pixels

Hamming distance

Page 24: Presentation project Computer Vision - Teeth segmentation

� � Created bitmap that contains most characteristics of

an incisor � If a segmented tooth can be described as a weighted

sum of a number of the bitmap images, it is classified as an incisor

Eigenfaces

Page 25: Presentation project Computer Vision - Teeth segmentation

� � Determine the principal components in both the

segmented image and the retrieved segmentation � Determine the distance between the principal

components

Principal Component Analysis (PCA)

Page 26: Presentation project Computer Vision - Teeth segmentation

� Questions?