An optical flow approach to tracking colonoscopy video

Liu, Jianfei; Subramanian, Kalpathi; Yoo, Terry S.

doi:10.1016/j.compmedimag.2013.01.010

Cited by 28 publications

(22 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…if K G = ∅, then the frame is labeled as normal. If one or more features satisfy (22), then we continue to the next step, where we compute a parameter upon which we base the decision whether the frame is classified as containing polyps or not.…”

Section: Geometrical Processing and The Tensor Of Intertiamentioning

confidence: 99%

“…we expect the polyps to be the protrusions that are somewhat rounded. Note that the combined geometric criterion (22) only uses the twodimensional information in u by only working with the binary segmentation s. To utilize information about the height of the protrusions, we need to work with u itself, which is why we fit the ball to u instead of s.…”

Section: E Decision Parameter and Binary Classifiermentioning

confidence: 99%

“…Finally, we can define the decision parameter by taking the maximum of the optimal fit ball radii over all the features that pass the combined geometric criterion (22) as…”

Section: E Decision Parameter and Binary Classifiermentioning

confidence: 99%

See 2 more Smart Citations

Automated Polyp Detection in Colon Capsule Endoscopy

Mamonov

Figueiredo

et al. 2014

IEEE Trans. Med. Imaging

216

105

View full text Add to dashboard Cite

Colorectal polyps are important precursors to colon cancer, a major health problem. Colon capsule endoscopy is a safe and minimally invasive examination procedure, in which the images of the intestine are obtained via digital cameras on board of a small capsule ingested by a patient. The video sequence is then analyzed for the presence of polyps. We propose an algorithm that relieves the labor of a human operator analyzing the frames in the video sequence. The algorithm acts as a binary classifier, which labels the frame as either containing polyps or not, based on the geometrical analysis and the texture content of the frame.We assume that the polyps are characterized as protrusions that are mostly round in shape. Thus, a best fit ball radius is used as a decision parameter of the classifier. We present a statistical performance evaluation of our approach on a data set containing over 18 900 frames from the endoscopic video sequences of five adult patients. The algorithm achieves 47% sensitivity per frame and 81% sensitivity per polyp at a specificity level of 90%. On average, with a video sequence length of 3747 frames, only 367 false positive frames need to be inspected by an operator.

show abstract

Section: Geometrical Processing and The Tensor Of Intertiamentioning

confidence: 99%

Section: E Decision Parameter and Binary Classifiermentioning

confidence: 99%

See 1 more Smart Citation

Automated Polyp Detection in Colon Capsule Endoscopy

Mamonov

Figueiredo

et al. 2014

IEEE Trans. Med. Imaging

216

105

View full text Add to dashboard Cite

show abstract

“…Otherwise, their optical flow algorithm relies on the fold structures of the colon which can be misleading due to is tubular structure and repeating fold shapes [8]. Furthermore, the colon is highly flexible in the sigmoid and transverse colon and here, Liu et al [9] describe problems when tracking the colonoscope due the large deformations and "sharp turns" of the colonoscope when it passes these regions. A continuous tracking of the colonoscope using optical flow proved very challenging in these regions of the colon and therefore several re-initializations of the method would need to take place.…”

Section: Related Workmentioning

confidence: 99%

Computer-assisted polyp matching between optical colonoscopy and CT colonography: a phantom study

et al. 2014

View full text Add to dashboard Cite

Potentially precancerous polyps detected with CT colonography (CTC) need to be removed subsequently, using an optical colonoscope (OC). Due to large colonic deformations induced by the colonoscope, even very experienced colonoscopists find it difficult to pinpoint the exact location of the colonoscope tip in relation to polyps reported on CTC. This can cause unduly prolonged OC examinations that are stressful for the patient, colonoscopist and supporting staff.We developed a method, based on monocular 3D reconstruction from OC images, that automatically matches polyps observed in OC with polyps reported on prior CTC. A matching cost is computed, using rigid point-based registration between surface point clouds extracted from both modalities. A 3D printed and painted phantom of a 25 cm long transverse colon segment was used to validate the method on two medium sized polyps. Results indicate that the matching cost is smaller at the correct corresponding polyp between OC and CTC: the value is 3.9 times higher at the incorrect polyp, comparing the correct match between polyps to the incorrect match. Furthermore, we evaluate the matching of the reconstructed polyp from OC with other colonic endoluminal surface structures such as haustral folds and show that there is a minimum at the correct polyp from CTC.Automated matching between polyps observed at OC and prior CTC would facilitate the biopsy or removal of true-positive pathology or exclusion of false-positive CTC findings, and would reduce colonoscopy false-negative (missed) polyps. Ultimately, such a method might reduce healthcare costs, patient inconvenience and discomfort.

show abstract

“…At present, the main methods of moving object extraction include optical flow method, the adjacent frame difference method, and the background subtraction method [2]. The optical flow method is based on the pixels' vector characteristics of image sequences for motion region detection, and it's accuracy for detection is affected by target characteristics, light illumination, target's velocity and noise and it's not suitable for the extraction of fragments from static explosion image sequences [3]. Neighbor frame difference method can extract the foreground of moving object by the gray difference of adjacent frame on image sequences which is insensitive to illumination change and noise interference [4], so it can be used to the extraction of the fragments.…”

Section: Introductionmentioning

confidence: 99%

Fragments Extraction and Segmentation Based on Three Frames Difference and Modified Otsu

Liu¹,

Zhang²,

Zhao³

2016

Proceedings of the 2015 5th International Conference on Computer Sciences and Automation Engineering

View full text Add to dashboard Cite

Abstract. This paper present a method for the extraction and segmentation of small targets in image sequence captured in static explosion test, whose background is complex with numerous interference. Three frames difference method is adopted to process the original image sequence to obtain the foreground image. Since the foreground image is heavily contaminated by noise, so the gray square enhancement method is used to increase the contrast between the noise and the target. The modified Otsu method is employed to obtain the global optimum threshold, and the segmentation of the foreground image is accomplished. Finally, morphological processing is performed, and the noise of the isolated points is eliminated, and the weakening of the target in the process of differential and segmentation is resumed. Through a series of processing, the extraction of fragments in static explosion test image sequence is completed, and the trajectories of the fragments are obtained.

show abstract

An optical flow approach to tracking colonoscopy video

Cited by 28 publications

References 30 publications

Automated Polyp Detection in Colon Capsule Endoscopy

Automated Polyp Detection in Colon Capsule Endoscopy

Computer-assisted polyp matching between optical colonoscopy and CT colonography: a phantom study

Fragments Extraction and Segmentation Based on Three Frames Difference and Modified Otsu

Contact Info

Product

Resources

About