Computer-aided tumor detection in endoscopic video using color wavelet features
read more
Citations
Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?
WM-DOVA maps for accurate polyp highlighting in colonoscopy: Validation vs. saliency maps from physicians
Automated Polyp Detection in Colonoscopy Videos Using Shape and Context Information
Towards automatic polyp detection with a polyp appearance model
Comparative Validation of Polyp Detection Methods in Video Colonoscopy: Results From the MICCAI 2015 Endoscopic Vision Challenge
References
Textural Features for Image Classification
A theory for multiresolution signal decomposition: the wavelet representation
A genetic model for colorectal tumorigenesis
Rank correlation methods
Genetic alterations during colorectal-tumor development.
Related Papers (5)
Frequently Asked Questions (12)
Q2. What have the authors stated for future works in "Computer-aided tumor detection in endoscopic video using color wavelet features" ?
Future extension of this work will be to determine a more robust classification scheme.
Q3. What is the purpose of the wavelet transform?
The wavelet transform offers a tool for spatial/frequency representation by decomposing the original images to the corresponding scales.
Q4. how many matrices are used to estimate the intensities of the color channels?
The cooccurrence matrices are estimated in four different directions of intensities’ relation, 0 , 45 , 90 , and 135 , resulting to 36 matrices(11)Finally the four statistical measures, namely angular second moment, correlation, inverse difference moment, and entropy are estimated for each matrix resulting in 144 wavelet features(12)where is the respective statistical measure.
Q5. How did the use of discrete wavelet transform affect the classification performance?
The use of discrete wavelet frame transform contributed to a significant increase of the classification performance by a factor of 2.4% and 12.2% to the values of specificity and sensitivity.
Q6. What are the three measures that derive from the cooccurrence matrix?
angular second moment, correlation, inverse difference moment and entropy(2)(3)(4)(5)where is the th entry of the normalized cooccurrence matrix, is the number of gray levels of the image, and and are the means and standard deviations of the marginal probability obtained by summing up the rows of matrix .
Q7. What is the technique for the detection of colorectal polyps?
A competitive new generation technique used for the detection of colorectal polyps is virtual colonoscopy based on computer tomography (CT) or magnetic resonance (MR) data.
Q8. What is the effect of the algorithm on the detection of colorectal polyps?
It has been applied on the detection of colorectal polyps in colonoscopic video frame sequences, and it has been found that the feature subspaces corresponding to normal and abnormal tissue are highly discriminant.
Q9. What was the optimal window size for the detection of polyps?
1) Second-Order Statistics on the Wavelet Domain of Grayscale Endoscopic Video Frames: Primarily, the color video frames were transformed to eight-bit intensity maps and the optimal window size for the detection of polyps was investigated.
Q10. What was the average classification performance of the second-order statistical features?
The second-order statistical features (2)–(5) were calculated directly from the intensity values of the corresponding windows, and the average classification performance was estimated % and % in terms of specificity and sensitivity, respectively.
Q11. What is the number of gray level features used?
According to the second-order statistics on the wavelet domain methodology and (2)–(7), the total number of the gray level features used is 72 (six cooccurrence measures 3 wavelet bands 4 directions).
Q12. What is the correlation threshold for the CWC feature set?
The fact that the sensitivity at 90% correlation threshold falls within the uncertainty range of the complete set of features (100% correlation threshold), suggests that the first four features (Table II) can be omitted, leading to the reduction of the feature space dimension by nine features, without any harmful implication in the resulted overall sensitivity.