Comparison of k-NN, SVM, and NN in Pit Pattern Classification of Zoom-Endoscopic Colon Images using Co-Occurrence Histograms

Häfner, Michael; Gangl, Alfred; Wrba, Friedrich; Thonhauser, K.; Schmidt, Holly; Kastinger, C.; Uhl, Andreas; Vécsei, Andreas

doi:10.1109/ispa.2007.4383747

Cited by 6 publications

(7 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…5, second row), from gray-level co-occurrence matrices, (Fig. 5, row three) [22][23][24][25][26][27][28][29], sum-and different histograms [20,21], texture spectrum [10], or local binary patterns [20,21,[30][31][32]. Examples for structural approaches were statistical-geometrical features [20] or features obtained from fractal analysis [29].…”

Section: Technical View: Statistical Description Of Texturesmentioning

confidence: 99%

“…Thirdly, starting in the mid-1990s, more sophisticated machine learning algorithms (than just simple thresholding) were available, which all have extensively been explored and applied for the aforementioned polyp-detection experiments based on the extracted texture features, such as “Bayes classifiers” [33, 34], “linear discriminators” [29], “nearest neighbor classifier” [20, 28-30, 39, 40], “artificial neural networks” (ANNs) [22-24, 26-28, 32, 36, 41, 42], “support vector machines” (SVMs) [10, 21, 28, 30, 31, 39, 40], and “random forest classifiers” [43].…”

Section: The Age Of Statistical Learningmentioning

confidence: 99%

See 1 more Smart Citation

Artificial Intelligence-Based Polyp Detection in Colonoscopy: Where Have We Been, Where Do We Stand, and Where Are We Headed?

Wittenberg

Raithel

2020

Visc Med

View full text Add to dashboard Cite

Background: In the past, image-based computer-assisted diagnosis and detection systems have been driven mainly from the field of radiology, and more specifically mammography. Nevertheless, with the availability of large image data collections (known as the “Big Data” phenomenon) in correlation with developments from the domain of artificial intelligence (AI) and particularly so-called deep convolutional neural networks, computer-assisted detection of adenomas and polyps in real-time during screening colonoscopy has become feasible. Summary: With respect to these developments, the scope of this contribution is to provide a brief overview about the evolution of AI-based detection of adenomas and polyps during colonoscopy of the past 35 years, starting with the age of “handcrafted geometrical features” together with simple classification schemes, over the development and use of “texture-based features” and machine learning approaches, and ending with current developments in the field of deep learning using convolutional neural networks. In parallel, the need and necessity of large-scale clinical data will be discussed in order to develop such methods, up to commercially available AI products for automated detection of polyps (adenoma and benign neoplastic lesions). Finally, a short view into the future is made regarding further possibilities of AI methods within colonoscopy. Key Messages: Researchofimage-based lesion detection in colonoscopy data has a 35-year-old history. Milestones such as the Paris nomenclature, texture features, big data, and deep learning were essential for the development and availability of commercial AI-based systems for polyp detection.

show abstract

Section: Technical View: Statistical Description Of Texturesmentioning

confidence: 99%

Section: The Age Of Statistical Learningmentioning

confidence: 99%

Artificial Intelligence-Based Polyp Detection in Colonoscopy: Where Have We Been, Where Do We Stand, and Where Are We Headed?

Wittenberg

Raithel

2020

Visc Med

View full text Add to dashboard Cite

show abstract

“…Analyses based on Kudo's classification have been reported using images mainly from magnifying chromoendoscopy (7,10,(12)(13)(14)(15)(16)(17) or narrow band imaging (NBI) endoscopy (18)(19)(20)(21)(22)(23)(24) and optical projection tomography (OPT) (25) . Most of them are performed by experienced endoscopists while a few used image software processing methods (4,12) .…”

Section: Ag-2019-50mentioning

confidence: 99%

Morphological Characterization of Colorectal Pits Using Autofluorescence Microscopy Images

Erbes

Zeitoune

Torres

et al. 2019

Arq. Gastroenterol.

View full text Add to dashboard Cite

BACKGROUND: Colorectal cancer is one of the most prevalent pathologies. Its prognosis is linked to the early detection and treatment. Currently diagnosis is performed by histological analysis from polyp biopsies, followed by morphological classification. Kudo’s pit pattern classification is frequently used for the differentiation of neoplastic colorectal lesions using hematoxylin-eosin stained samples. Few articles have reported this classification with image software processing, using exogenous markers over the samples. The processing of autofluorescence images is an alternative that could allow the characterization of the pits from the crypts of Lieberkühn, bypassing staining techniques. OBJECTIVE: Processing and analysis of widefield autofluorescence microscopy images obtained by fresh colon tissue samples from a murine model of colorectal cancer in order to quantify and characterize the pits morphology by measuring morphology parameters and shape descriptors. METHODS: Adult male BALB/cCmedc strain mice (n=27), ranging from 20 to 30 g, were randomly assigned to four and five groups of treated and control animals. Colon samples were collected at day zero and at fourth, eighth, sixteenth and twentieth weeks after treatmentwith azoxymethane. Two-dimensional (2D) segmentation, quantification and morphological characterization of pits by image processing applied using macro programming from FIJI. RESULTS: Type I is the pit morphology prevailing between 53 and 81% in control group weeks. III-L and III-S types were detected in reduced percentages. Between the 33 and 56% of type I was stated as the prevailing morphology for the 4th, 8th and 20th weeks of treated groups, followed by III-L type. For the 16th week, the 39% of the pits was characterized as III-L type, followed by type I. Further, pattern types as IV, III-S and II were also found mainly in that order for almost all of the treated weeks. CONCLUSION: These preliminaries outcomes could be considered an advance in two-dimensional pit characterization as the whole image processing, comparing to the conventional procedure, takes a few seconds to quantify and characterize non-pathological colon pits as well as to estimate early pathological stages of colorectal cancer.

show abstract

“…Similar structured patterns (co-called pit patterns) can be observed on the walls of colon, which are known to relate to different stages of dysplastic tissue. To describe and discriminate these pit patterns, Heafner et al [5,6] have sug-…”

Section: State Of the Artmentioning

confidence: 99%

“…All feature vectors have been calculated independently from different color planes including RGB, HSI and YCrCb. For the classification of the feature vectors, neural networks, the k-nearest neighbor as well as support vector machines (SVM) [6] and the Bayes classifier [5] have been proposed. On a database with 198 non-neoplastic and 286 neoplastic cases, the best results were achieved using the Bayes classifier yielding a mean classification rate of 96% on a two-class problem with optimized ring filters in the power spectrum calculated in the color planes of the RGB color space.…”

Section: Overviewmentioning

confidence: 99%

Computer-assisted Diagnosis for Precancerous Lesions in the Esophagus

Münzenmayer¹,

Kage²,

Wittenberg³

et al. 2009

Methods Inf Med

View full text Add to dashboard Cite

show abstract

Comparison of k-NN, SVM, and NN in Pit Pattern Classification of Zoom-Endoscopic Colon Images using Co-Occurrence Histograms

Abstract: Abstract

Cited by 6 publications

References 16 publications

Artificial Intelligence-Based Polyp Detection in Colonoscopy: Where Have We Been, Where Do We Stand, and Where Are We Headed?

Artificial Intelligence-Based Polyp Detection in Colonoscopy: Where Have We Been, Where Do We Stand, and Where Are We Headed?

Morphological Characterization of Colorectal Pits Using Autofluorescence Microscopy Images

Computer-assisted Diagnosis for Precancerous Lesions in the Esophagus

Contact Info

Product

Resources

About