Chapter 7 Flashcards
What does BLOB stand for?
Binary large object.
What is connectivity
Relates to whether to objects are connected or not.
Two types of connectivity:
4- and 8-connectivity.
Grass-fire:
Pretend you light one place on fire (top left) and that the fire can only travel via 4-or 8-connectivity.
What are BLOB-features?
Different attributes of BLOBs -this could be circularity. Often held up against known object.
BLOB area:
Number of pixels in a blob
BLOB bounding box:
The quadrate around the blob with sides lowest x and highest x and top and bottom lowest y and highest y.
Bounding box ratio:
Height of box divided by width.
Bounding circle:
same as BB but a circle.Found by finding center then looking in all directions.
Feature vector:
All BLOB features collected in a vector.
Convex hull:
minimum novex ploygon.
BLOB compactness:
Area/BB_area
Center of mass:
(average_x_coordinate, average_y_coordinate)
Perimeter:
Like f(x,y) - erosion f(x,y)
Circularity:
Perimeter_measured/perimeter_estimated
BLOB classification
BLOBs can be classified on the basis of their features being within certain values.
Feature space:
For Blob features a feature space can be made -basicallye every axis is a feature and you must stay within certain values of each feature to be in the feature space.
What is a confusion matrix?
matrix listing
TN FP
FN TP
Accuracy:
(TP+TN)/N
Misclassification rate:
(FN+FP)/N
True positive rate / sensitivity:
TP / (TP + FN)
False positive rate
FP / (TN + FP)
Specificity
TN / (TN + FP)
Precision
TP / (FP + TP)
