Automatic Minirhizotron Root Image Analysis

1
Automatic Minirhizotron Root Image Analysis
Using Two-Dimensional Matched Filtering and
Local Entropy Thresholding
Presented by Guang Zeng
2
Importance of Studying Roots
3
Methods for Studying Roots
Minirhizotron
Soil Sampling
Rhizotron
4
Previous work on minirhizotron image analysis

• Vamerali Ganis 1999
• Nonlinear contrast stretching technique is
  used to enhance the local contrast of
  rootsLimitation The minimum root length filter
  will eliminate some shorter roots.
• Natar Baker 1992
• An artificial neural system is developed to
  identify rootsLimitation The accuracy will
  substantial decrease when applied to images that
  have not been trained.
• Dowdy Smucker 1998The length-to-diameter
  ratio is used to discriminate roots Limitation
  Only works for a single type of root.

5
Preview of Experimental Results
Original image
Extracted root
Measured root
6
Approach Overview
7
Image Preprocessing
1. Conversion to grayscale
Red
Green
Blue
Color
2. Contrast stretching
We set
3. Smoothing the image
8
Matched Filtering Principles
Similarity between plant roots and blood vessels

• Small curvature
• Parallel edges

? piecewise linear segmentsChaidhuri et. 1989
Motivation

• Young roots appear brighter
• Gaussian curve for gray level profile of
• cross section

9
Matched Filtering Procedure

• A number of cross sections of identical profiles
  are matched
• simultaneously. A kernel can be used which
  mathematically
• expressed as

for y L/2
where L is the length of the segment for which
the root is assumed to have a fixed orientation.

• Kernels, for which the mean value is positive,
  are forced to
• have slightly negative mean values in order to
  reduce the effect
• of background noise.

10
Matched Filtering Procedure (cont.)

• The kernel is rotated using an angular
  resolution of 15 (12 kernels are needed to
  span all possible orientations).

(a) 15
(b) 75
(c) 135
(d) 180

• The kernel is applied at two scales (full
  image size and half image size, obtained by
  subsampling).

11
Matched Filtering Output
(a) 75
(b) 90
(c) 135
(d) 180
12
Local Entropy Thresholding
Shannons entropy
and
where,
13
Local Entropy Thresholding (cont.)
The probability of co-occurrence pij of gray
levels i and j can therefore be written as
Divide co-occurrence matrix into quadrants, using
threshold t (0 t L)
The local entropy is defined by the quadrants A
and D.
14
Local Entropy Thresholding (cont.)
Background-to-background entropy
Foreground-to-foreground entropy
Hence, the total second-order local entropy of
the object and the background can be written
as The gray level corresponding to the maximum
of HT(t) gives the optimal threshold for
object-background classification.
15
Local Entropy Thresholding Outputs
(a) 75
(b) 90
(c) 135
(d) 180
t 122
t 103
t 155
t 130
16
Selecting the Root
1. Connected component labeling
2. Root candidate selecting (Ai 0.8 Amax )
17
Comparison of Root Selection Methods
Chanwinmaluang and Fan 2003
Our method
originalimage
...
separate MF outputs
combined MF output
detected root
detected root
18
Root Measurement
1. Object Skeletonization
2. Extracting medial line using Dijkstras
Algorithm
19
Root Measurement (cont.)
3. Estimating the length
Freeman formula
Pythagorean theorem
Kimuras method
20
Root Measurement (cont.)
4. Estimating the average diameter
Step 1 Select 10 nodes that equally divide the
medial line into 11 parts.
Step 2 Find the corresponding opposite boundary
point pairs, calculate the distance between
each opposite boundary point pairs.
Step 3 Discard the two pairs that yield the
maximum and the minimum distance
21
Root Discrimination
False positives are caused by
1. A bright extraneous object
2. Uneven diffusion of light through the
minirhizotron wall
22
Root Discrimination Five Methods
2. Approximate line symmetry
3. Boundary parallelism
1. Eccentricity
e c / a
23
Root Discrimination Five Methods (cont.)
4. Histogram Distribution
5. Edge Detection
24
Experimental Results

• We tested our method on a set of 45
  minirhizotron images containing
• different sizes of roots
• different types of roots
• dead roots
• no roots

• The output of the algorithm is compared with
  hand-labeled
• ground truth provided by the Clemson Root
  Biology Lab.

25
Experimental Results (cont.)
Original image
Extracted root
Measured root
26
Experimental Results (cont.)
Original image
Extracted root
Measured root
27
Experimental Results (cont.)
Original image
Extracted root
Measured root
28
Comparison of Root Length Measurement Methods
1. Measurement Deviation
2. Correlation
29
Comparison of Root Discrimination Methods
1. The optimal threshold point is the closest
point to the perfect result. The closer the
optimal threshold point to the point (0,1), the
more accurate the method. 2. The larger the area
beneath an ROC curve, the more accurate the
method.
30
Multiple root detection

• Our technique is limited to zero or one root per
  image.
• We tried detecting multiple roots by extracting
  the two largest components in the thresholded
  binary images, then running our algorithm.
• Some results

• Works on some images, but the false positive
  rate is increased to 14 (more bright background
  objects are misclassified).

31
Conclusion

• Fully automatic algorithm for detecting and
  measuring roots
• Works on multiple root types
• Uses individual matched filters outputs, without
  first combining
• them.
• Uses a robust thresholding method
• Robust medial line detection using Dijkstras
  algorithm
• Proposed five different methods for root /
  no-root discrimination

Future work
1. Accurate multi-root detection
2. Reducing the computation time