A-DenseUNet: Adaptive Densely Connected UNet for Polyp Segmentation in Colonoscopy Images with Atrous Convolution

Safarov, Sirojbek; Whangbo, Taeg Keun

doi:10.21203/rs.3.rs-158417/v1

Cited by 8 publications

(6 citation statements)

References 36 publications

(67 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Models based on dilated convolution architecture: Sun et al 58 used dilated convolution in the last block of the encoder while Safarov et al 59 used in all encoder blocks. Though 59 used a mesh of attention blocks and residual block as a decoder, both methods tested there model on CVC-ClinicDB achieving F1-score of 96.106 and 96.043, respectively.…”

Section: /26mentioning

confidence: 99%

Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge

Ali¹,

Ghatwary²,

Jha³

et al. 2022

Preprint

View full text Add to dashboard Cite

Polyps are well-known cancer precursors identified by colonoscopy. However, variability in their size, location, and surface largely affect identification, localisation, and characterisation. Moreover, colonoscopic surveillance and removal of polyps (referred to as polypectomy ) are highly operator-dependent procedures. There exist a high missed detection rate and incomplete removal of colonic polyps due to their variable nature, the difficulties to delineate the abnormality, the high recurrence rates, and the anatomical topography of the colon. There have been several developments in realising automated methods for both detection and segmentation of these polyps using machine learning. However, the major drawback in most of these methods is their ability to generalise to out-of-sample unseen datasets that come from different centres, modalities and acquisition systems. To test this hypothesis rigorously we curated a multi-centre and multi-population dataset acquired from multiple colonoscopy systems and challenged teams comprising machine learning experts to develop robust automated detection and segmentation methods as part of our crowd-sourcing Endoscopic computer vision challenge (EndoCV) 2021. In this paper, we analyse the detection results of the four top (among seven) teams and the segmentation results of the five top teams (among 16). Our analyses demonstrate that the top-ranking teams concentrated on accuracy (i.e., accuracy > 80% on overall Dice score on different validation sets) over real-time performance required for clinical applicability. We further dissect the methods and provide an experiment-based hypothesis that reveals the need for improved generalisability to tackle diversity present in multi-centre datasets.Author contributions S. Ali conceptualised the work, led the challenge and workshop, prepared the dataset, software and performed all analyses. S. Ali, N. Ghatwary and D. Jha contributed in data annotations. T. de Lange, J.E. East, S. Realdon, R. Cannizzaro, D. Lamarque were involved providing colonoscopy data and in the validation and quality checks of the annotations used in this challenge. Challenge participants (E

show abstract

Section: /26mentioning

confidence: 99%

Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge

Ali¹,

Ghatwary²,

Jha³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…These networks are also widely used for segmentation and classification tasks. Wanwan Zhang et al [16] proposed a kind of densely connected U-Net with extended convolution and attentional mechanisms. However, obtaining global information is a difficult task, which becomes more difficult when data sets are complex.…”

Section: Related Workmentioning

confidence: 99%

Brain Tumor Image Segmentation Based on Grouped Convolution

Wang

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

The automatic segmentation of MRI multi-modal images of brain tumors is one of the important research contents of disease detection and analysis. Due to the heterogeneity of tumors, it is difficult to achieve efficient and accurate automatic segmentation of brain tumors. Traditional segmentation methods based on machine learning cannot handle complex scenes such as complex edges and overlapping categories. In clinical assisted diagnosis, it is of great significance to apply deep learning to two-dimensional natural image segmentation and three-dimensional medical image segmentation. In this paper, we propose a three-dimensional network model to achieve precise segmentation of brain tumors. The model adopts an encoder-decoder structure and replaces ordinary convolution with grouped convolution to reduce network parameters and improve network performance. The model improves the problem of information exchange between different groups through channel mixing. Experiments conducted on the BraTS (Brain Tumor Segmentation) 2018 challenge dataset prove that our network greatly reduces the computational cost while ensuring segmentation accuracy.

show abstract

“…Authors in [9] proposed an automatic GP segmentation based on HOG feature map and saliency enhancement, in [10] authors implemented a joint thresholding and segmentation of GP employed bran-storm-optimization supported Kapur's thresholding to enhance the image. In [11] authors proposed a CNN supported semantic-segmentation methodology to localize the GP from the CI, in [12] authors implemented CNN supported A-DenseUNet to extract and examine the GP fragment of CVC and Kvasir separately. Authors in [13] proposed MED-Net to separately evaluates CVC and ETIS datasets.…”

Section: Contextmentioning

confidence: 99%

“…In this phase, the β is chosen as 1.5 and other values, which helps to get the spiral search is given in Eqs. ( 8)- (12).…”

Section: Aquila-optimization-algorithmmentioning

confidence: 99%

Semi/Fully-Automated Segmentation of Gastric-Polyp Using Aquila-Optimization-Algorithm Enhanced Images

Rajinikanth¹,

Aslam²,

Kadry³

et al. 2022

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

The incident rate of the Gastrointestinal-Disease (GD) in humans is gradually rising due to a variety of reasons and the Endoscopic/Colonoscopic-Image (EI/CI) supported evaluation of the GD is an approved practice. Extraction and evaluation of the suspicious section of the EI/CI is essential to diagnose the disease and its severity. The proposed research aims to implement a joint thresholding and segmentation framework to extract the Gastric-Polyp (GP) with better accuracy. The proposed GP detection system consist; (i) Enhancement of GP region using Aquila-Optimization-Algorithm supported tri-level thresholding with entropy (Fuzzy/Shannon/Kapur) and between-class-variance (Otsu) technique, (ii) Automated (Watershed/Markov-Random-Field) and semi-automated (Chan-Vese/Level-Set/Active-Contour) segmentation of GP fragment, and (iii) Performance evaluation and validation of the proposed scheme. The experimental investigation was performed using four benchmark EI dataset (CVC-ClinicDB, ETIS-Larib, EndoCV2020 and Kvasir). The similarity measures, such as Jaccard, Dice, accuracy, precision, sensitivity and specificity are computed to confirm the clinical significance of the proposed work. The outcome of this research confirms that the fuzzyentropy thresholding combined with Chan-Vese helps to achieve a better similarity measures compared to the alternative schemes considered in this research.

show abstract

A-DenseUNet: Adaptive Densely Connected UNet for Polyp Segmentation in Colonoscopy Images with Atrous Convolution

Cited by 8 publications

References 36 publications

Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge

Assessing generalisability of deep learning-based polyp detection and segmentation methods through a computer vision challenge

Brain Tumor Image Segmentation Based on Grouped Convolution

Semi/Fully-Automated Segmentation of Gastric-Polyp Using Aquila-Optimization-Algorithm Enhanced Images

Contact Info

Product

Resources

About