Abstract
Manual detection of colorectal polyps in the colonoscopy videos during diagnosis of colorectal cancer can be challenging due to the large no. of sequential frames and high chances of false positive rate. So, researchers are working to develop an automatic computer-aided decision system to successfully detect colorectal polyps for early diagnosis of colorectal cancer. This work proposes an end-to-end, automatic colorectal polyp detection architecture called ‘Window-based Detection after Mixed Convolutions Polyp Identification (WD-MCPI)’. It has been trained and validated on the Etis-Larib, CVC-Colon, and Kvasir v1 data. Varying sequential and non-sequential colonoscopy frames and developed occlusive frames have been considered in the test set. Systematic hyperparameter tuning of various convolutional layers, optimizers, kernel size, color space, image dimension, and filter size has been performed to build the proposed architecture. The robustness and explainability of the proposed architecture have been estimated through various evaluation metrics, feature mapping, ablation studies, occlusion testing, and class activation mapping. The proposed work has achieved an overall accuracy, precision, recall, specificity, F1 score, and area-under-curve score up to 94.23, 91.16, 94.00, 92.67, 91.75, and 92.53%, respectively. An average test set accuracy up to 93% has been achieved on a newly developed test set named ‘Gastrointestinal atlas-Colon Polyp’ (available here). Upon comparison with the existing state-of-the-art works and vanilla inception v3 architecture, the proposed architecture has been tested on varying colonoscopy frames, achieving optimum results with reduced trainable parameters, and inference time per epoch.
Similar content being viewed by others
Data availability
The datasets used in this work are freely available online. Refer link.
Code availability
Available on request from the corresponding author.
References
Shah D, Makharia GK, Ghoshal UC, Varma S, Ahuja V, Hutfless S (2018) Burden of gastrointestinal and liver diseases in india, 1990–2016. Indian J Gastroenterol 37(5):439–445
Mohandas KM (2011) Colorectal cancer in India: controversies, enigmas and primary prevention. Indian J Gastroenterol 30(1):3–6
Stryker SJ, Wolff BG, Culp CE, Libbe SD, Ilstrup DM, MacCarty RL (1987) Natural history of untreated colonic polyps. Gastroenterology 93(5):1009–1013
Oh J, Hwang S, Cao Y, Tavanapong W, Liu D, Wong J, De Groen PC (2008) Measuring objective quality of colonoscopy. IEEE Trans Biomed Eng 56(9):2190–2196
Chan MY, Cohen H, Spiegel BM (2009) Fewer polyps detected by colonoscopy as the day progresses at a veteran’s administration teaching hospital. Clin Gastroenterol Hepatol 7(11):1217–1223
Goel N, Kaur S, Gunjan D, Mahapatra S (2022) Investigating the significance of color space for abnormality detection in wireless capsule endoscopy images. Biomed Signal Process Control 75:103624
Nogueira-Rodriguez A, Dominguez-Carbajales R, Lopez-Fernandez H, Igle-sias A, Cubiella J, Fdez-Riverola F, Reboiro-Jato M, Glez-Pena D (2021) Deep neural networks approaches for detecting and classifying colorectal polyps. Neurocomputing 423:721–734
Goel N, Kaur S, Gunjan D, Mahapatra S (2022) Dilated cnn for abnormality detection in wireless capsule endoscopy images. Soft Comput 26(3):1231–1247
Ikematsu H, Murano T, Shinmura K (2022) Detection of colorectal lesions during colonoscopy. DEN open 2(1):68
Houwen BB, Nass KJ, Vleugels JL, Fockens P, Hazewinkel Y, Dekker E (2022) Comprehensive review of publicly available colonoscopic imaging databases for artificial intelligence research: availability, accessibility and usability. Gastrointest Endosc 97(2):184–199
Wickstrøm K, Kampffmeyer M, Jenssen R (2020) Uncertainty and interpretability in convolutional neural networks for semantic segmentation of colorectal polyps. Med Image Anal 60:101619
Hicks S, Storas A, Riegler M, Midoglu C, Hammou M, Lange T, Parasa S, Halvorsen P, Strumke I (2022) Visual explanations for polyp detection: How medical doctors assess intrinsic versus extrinsic explanations. arXiv preprint arXiv:2204.00617
Chen J, Li K, Rong H, Bilal K, Yang N, Li K (2018) A disease diagnosis and treatment recommendation system based on big data mining and cloud computing. Inf Sci 435:124–149
Chen J, Yang N, Zhou M, Zhang Z, Yang X (2022) A configurable deep learning framework for medical image analysis. Neural Comput Appl 34(10):7375–7392
Zhao L, Li N, Tan G, Chen J, Li S, Duan M (2022) The end-to-end fetal head circumference detection and estimation in ultrasound images. IEEE/ACM Trans Comput Biol Bioinf. https://doi.org/10.1109/TCBB.2022.3227037
Jie L, Liang P, Zhao Z, Chen J, Chang Q, Zeng Z (2022) Adan: an adversarial domain adaptation neural network for early gastric cancer prediction. In: 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pp. 2169–2172 (2022)
Handa P, Goel N, Indu S (2022) Automatic intestinal content classification using transfer learning architectures. In: 2022 IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT), pp. 1–5
ELKarazle K, Raman V, Then P, Chua C (2023) Detection of colorectal polyps from colonoscopy using machine learning: A survey on modern techniques. Sensors 23(3):1225
Shah S, Park N, Chehade NEH, Chahine A, Monachese M, Tiritilli A, Moosvi Z, Ortizo R, Samarasena J (2023) Effect of computer-aided colonoscopy on adenoma miss rates and polyp detection: a systematic review and meta-analysis. J Gastroenterol Hepatol 38(2):162–176
Hsu C-M, Hsu C-C, Hsu Z-M, Shih F-Y, Chang M-L, Chen T-H (2021) Colorectal polyp image detection and classification through grayscale images and deep learning. Sensors 21(18):5995
Akbari M, Mohrekesh M, Rafiei S, Soroushmehr SR, Karimi N, Samavi S, Najarian K (2018) Classification of informative frames in colonoscopy videos using convolutional neural networks with binarized weights. In: 2018 40th Annual Inter- national Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pp. 65–68
Rahim T, Hassan SA, Shin SY (2021) A deep convolutional neural network for the detection of polyps in colonoscopy images. Biomed Signal Process Control 68:102654
Gong R, He S, Tian T, Chen J, Hao Y, Qiao C (2023) FRCNN-AA-CIF: An automatic detection model of colon polyps based on attention awareness and context information fusion. Comput Biol Med 158:106787
Kara OC, Venkatayogi N, Ikoma N, Alambeigi F (2023) A reliable and sensitive framework for simultaneous type and stage detection of colorectal cancer polyps. Annals Biomed Eng 51(7):1–14
Chino A, Ide D, Abe S, Yoshinaga S, Ichimasa K, Kudo T, Ninomiya Y, Oka S, Tanaka S, Igarashi M (2023) Performance evaluation of a computer- aided polyp detection system with artificial intelligence for colonoscopy. Dig Endosc. https://doi.org/10.1111/den.14578
Younas F, Usman M, Yan WQ (2023) A deep ensemble learning method for colorectal polyp classification with optimized network parameters. Appl Intell 53(2):2410–2433
Mangotra H, Goel N et al (2023) Effect of selection bias on automatic colonoscopy polyp detection. Biomed Signal Process Control 85:104915
Wang S, Yin Y, Wang D, Lv Z, Wang Y, Jin Y (2021) An interpretable deep neu- ral network for colorectal polyp diagnosis under colonoscopy. Knowl Based Syst 234:107568
Kliegis L, Obst W, Bruns J, Weigt J (2022) Can a polyp detection and characterization system predict complete resection? Dig Dis 40(1):115–118
Kaiser L, Gomez AN, Chollet F (2017) Depthwise separable convolutions for neural machine translation. arXiv preprint arXiv:1706.03059
Dumoulin V, Visin F (2016) A guide to convolution arithmetic for deep learning. arXiv preprint arXiv:1603.07285
Kavukcuoglu K, Sermanet P, Boureau Y-L, Gregor K, Mathieu M, Cun Y et al. (2010) Learning convolutional feature hierarchies for visual recognition. Advances in neural information processing
Agarap AF (2018) Deep learning using rectified linear units (relu). arXiv preprint arXiv:1803.08375
Basta N (2022) The Differences between Sigmoid and Softmax Activation Func- tions, Medium. https://medium.com/arteos-ai/the-differences-between-sigmoid-and-softmax-acti\protect\discretionary{\char\hyphenchar\font}{}{}vation-function-12adee8cf322. Accessed 21 Sept 2022
Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D (2017) Grad-cam: Visual explanations from deep networks via gradient-based localiza- tion. In: Proceedings of the IEEE International Conference on Computer Vision, pp 618–626
Lafeuille P, Lambin T, Yzet C, Latif EH, Ordoqui N, Rivory J, Pioche M (2022) Flat colorectal sessile serrated polyp: an example of what artificial intelligence does not easily detect. Endoscopy 54(05):520–521
Yu T, Lin N, Zhong X, Zhang X, Zhang X, Chen Y, Liu J, Hu W, Duan H, Si J (2022) Multi-label recognition of cancer-related lesions with clinical priors on white-light endoscopy. Comput Biol Med 143:105255
Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z (2016) Rethinking the inception architecture for computer vision. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2818–2826
Patel K, Li K, Tao K, Wang Q, Bansal A, Rastogi A, Wang G (2020) A com-parative study on polyp classification using convolutional neural networks. PLoS ONE 15(7):0236452
Jheng Y-C, Wang Y-P, Lin H-E, Sung K-Y, Chu Y-C, Wang H-S, Jiang J-K, Hou M-C, Lee F-Y, Lu C-L (2022) A novel machine learning- based algorithm to identify and classify lesions and anatomical landmarks in colonoscopy images. Surg Endosc 36(1):640–650
Jia X, Mai X, Cui Y, Yuan Y, Xing X, Seo H, Xing L, Meng MQ-H (2020) Automatic polyp recognition in colonoscopy images using deep learning and two- stage pyramidal feature prediction. IEEE Trans Auto Sci Eng 17(3):1570–1584
Ellahyani A, Jaafari IE, Charfi S, Ansari ME (2022) Fine-tuned deep neural networks for polyp detection in colonoscopy images. Pers Ubiquit Comput 27(2):1–13
Karaman A, Pacal I, Basturk A, Akay B, Nalbantoglu U, Coskun S, Sahin O, Karaboga D (2023) Robust real-time polyp detection system design based on yolo algorithms by optimizing activation functions and hyper-parameters with artificial bee colony (abc). Expert Syst Appl 221:119741
Funding
No funding was received for this work.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare no personal, academic, or financial conflicts of interest associated with this paper.
Ethical approval
This article does not contain any studies with human participants or animals performed by any of the authors.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Handa, P., Goel, N., Indu, S. et al. Automatic Detection of Colorectal Polyps with Mixed Convolutions and its Occlusion Testing. Neural Comput & Applic 35, 19409–19426 (2023). https://doi.org/10.1007/s00521-023-08762-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-023-08762-z