NanoNet: Real-Time Polyp Segmentation in Video Capsule Endoscopy and Colonoscopy

Jha, Debesh; Tomar, Nikhil Kumar; Ali, Sharib; Riegler, Michael; Johansen, Håvard D.; Johansen, Dag; Lange, Thomas de; Halvorsen, Pål

doi:10.1109/cbms52027.2021.00014

Cited by 47 publications

(22 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Model IOU Dice coef Ours 0.8322 0.8967 Res-UNet++ [11] 0.7927 0.8133 DDA-Net [13] 0.8576 0.8201 Res-UNet++ +TGA+CRF [11] 0.7952 0.8508 Compete with other models such as Res-UNet [6], DDA-Net [13], Nano-Net [12], Res-UNet++ [11], our networks show great metric results on different datasets. This indicates that the model generalizes well to various datasets and validates the effectiveness of location embedding methods and multi-kernel-size convolution in segmentation tasks.…”

Section: )mentioning

confidence: 76%

“…Results on the Automatic Polyps dataset: The Automatic Polyps dataset [9] is derived from the Medico Automatic Polyp Segmentation Challenge aimed at developing a computer-aided diagnostic system for automated polyp segmentation to detect all types of polyps. As shown in table I, our model achieves very good results, comparing Res-UNet [6] and NaNo-Net-A [12] we exceeded the score. 3) Results on the EndoTect 2020 dataset: The EndoTect 2020 dataset [29] is introduced in the EndoTect Challenge.…”

Section: )mentioning

confidence: 78%

“…This challenge aims to facilitate the development of algorithms to help medical professionals uncover commonly occurring abnormalities in the gastrointestinal tract. Again, PEFNet performed surprisingly well with his IOU score of 0.7607 and Dice coefficient score of 0.8406, outperforming both DDA-Net [13] and NanoNet [12]. The results are shown in table II.…”

Section: )mentioning

confidence: 91%

“…A more effective version of Res-UNet architecture for colonoscopic image segmentation is called Res-UNet++ [11], which effectively fuses novel modules to improve the encoding and skip operations. A lightweight model with low latency that can be incorporated with low-end endoscope hardware is called Nano-Net [12]. DDA-Net [13], based on a dual decoder attention network, can generate the attention map to improve the semantic representation of the feature maps.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Multi Kernel Positional Embedding ConvNeXt for Polyp Segmentation

Nguyen-Mau

Trinh

Bui

et al. 2022

2022 RIVF International Conference on Computing and Communication Technologies (RIVF)

View full text Add to dashboard Cite

Medical image segmentation is the technique that helps doctor view and has a precise diagnosis, particularly in Colorectal Cancer. Specifically, with the increase in cases, the diagnosis and identification need to be faster and more accurate for many patients; in endoscopic images, the segmentation task has been vital to helping the doctor identify the position of the polyps or the ache in the system correctly. As a result, many efforts have been made to apply deep learning to automate polyp segmentation, mostly to ameliorate the U-shape structure. However, the simple skip connection scheme in UNet leads to the deficient context information and the semantic gap between feature maps from the encoder and decoder. To deal with this problem, we propose a novel framework composed of ConvNeXt backbone and Multi Kernel Positional Embedding block. Thanks to the suggested module, our method can attain better accuracy and generalization in the polyps segmentation task. Extensive experiments show that our model achieves the Dice coefficient of 0.8818 and the IOU score of 0.8163 on the Kvasir-SEG dataset. Furthermore, on various datasets, we make competitive achievement results with other previous state-of-the-art methods.

show abstract

Section: )mentioning

confidence: 76%

Section: )mentioning

confidence: 78%

Section: )mentioning

confidence: 91%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Multi Kernel Positional Embedding ConvNeXt for Polyp Segmentation

Nguyen-Mau

Trinh

Bui

et al. 2022

2022 RIVF International Conference on Computing and Communication Technologies (RIVF)

View full text Add to dashboard Cite

show abstract

“…We evaluated our model's performance with a benchmark consisting of the state-of-the-art models U-Net, DoubleU-Net, ResUNet++, NanoNet A [32], NanoNet B [32] and NanoNet C [32]. TableIV shows our model's performance results for Dice, IoU, AUC, precision and recall metrics compared to the results of the benchmark studies.…”

Section: Surgical Instrument Segmentationmentioning

confidence: 99%

PlutoNet: An Efficient Polyp Segmentation Network with Modified Partial Decoder and Decoder Consistency Training

Erol¹,

Sarıkaya²

2022

Preprint

View full text Add to dashboard Cite

Polyps in the colon can turn into cancerous cells if not removed with early intervention. Deep learning models are used to minimize the number of polyps that goes unnoticed by the experts, and to accurately segment the detected polyps during these interventions. Although these models perform well on these tasks, they require too many parameters, which can pose a problem with real-time applications. To address this problem, we propose a novel segmentation model called PlutoNet which requires only 2,626,337 parameters while outperforming state-of-the-art models on multiple medical image segmentation tasks. We use EfficientNetB0 architecture as a backbone and propose the novel modified partial decoder, which is a combination of partial decoder and full scale connections, which further reduces the number of parameters required, as well as captures semantic details. We use asymmetric convolutions to handle varying polyp sizes. Finally, we weight each feature map to improve segmentation by using a squeeze and excitation block. In addition to polyp segmentation in colonoscopy, we tested our model on segmentation of nuclei and surgical instruments to demonstrate its generalizability to different medical image segmentation tasks. Our model outperformed the state-of-the-art models with a Dice score of %92.3 in CVC-ClinicDB dataset and %89.3 in EndoScene dataset, a Dice score of %91.93 on the 2018 Data Science Bowl Challenge dataset, and a Dice score of %94.8 on Kvasir-Instrument dataset. Our experiments and ablation studies show that our model is superior in terms of accuracy, and it is able generalize well to multiple medical segmentation tasks.

show abstract