An automatic segmentation of endometrial cancer on ultrasound images

Thampi, Lidiya; Malarkhodi, S.

doi:10.1109/iccsp.2013.6577032

Cited by 5 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Such a key-point discriminator improved the baseline performance. On the other hand, Thampi et al [10] focused on the automatic segmentation of endometrial cancer from ultrasounds images, through level set and Otsu's thresholding methods. From a broader perspective, Usha et al [11] investigate the automatic measurement of the ovarian size and its shape parameters assessment to help experts to do a quick diagnosis, while Jin et al [12] analyze the accuracy of segmentation algorithms based on multiple U-net models, applied to ultrasound images, in patients with ovarian cancer.…”

Section: Related Workmentioning

confidence: 99%

Multi-Scale Deep Learning Ensemble for Segmentation of Endometriotic Lesions

Podda

Balia

Barra

et al. 2023

Preprint

View full text Add to dashboard Cite

Ultrasound is a readily available, non-invasive and low cost screening for the identification of endometriosis lesions, but its diagnostic specificity strongly depends on the experience of the operator. For this reason, computer-aided diagnosis tools based on Artificial Intelligence techniques can provide significant help to the clinical staff, both in terms of workload reduction and in increasing the overall accuracy of this type of examination and its outcome. However, although these techniques are spreading rapidly in a variety of domains, their application to endometriosis is still very limited. To fill this gap, we propose and evaluate a novel multi-scale ensemble approach for the automatic segmentation of endometriosis lesions from trans-vaginal ultrasounds. The peculiarity of the method lies in its high discrimination capability, obtained by combining, in a fusion fashion, multiple Convolutional Neural Networks trained on data at different granularity. The experimental validation carried out shows that: (i) the proposed method allows to significantly improve the performance of the individual neural networks, even in the presence of a limited training set; (ii) with a Dice coefficient of 82%, it represents a valid solution to increase the diagnostic efficacy of the ultrasound examination against such a pathology.

show abstract

Section: Related Workmentioning

confidence: 99%

Multi-Scale Deep Learning Ensemble for Segmentation of Endometriotic Lesions

Podda

Balia

Barra

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Performans karşılaştırmaları tam referans bazlı metrikler olan ortalama hata karesi (MSE), ortalama karesel hata karekökü (RMSE), SNR ve tepe sinyal gürültü oranı (PSNR) gibi istatistiksel ölçütler ile yapılmıştır. Diğer bir çalışmada endometrium kanseri segmentasyonu işleminden önce ultrason görüntüsü çeşitli filtreler ile filtrelenmiş ve en iyi sonucu SRAD filtresi vermiştir [5] fakat filtre performansları değerlendirilirken yine tam referans bazlı metrikler kullanılmıştır. Başka bir çalışmada [6], meme lezyonlarının ultrason görüntüleri üzerinde benek gürültüleri giderme filtrelerinin karşılaştırılması yapılmış, referans görüntüye ihtiyaç duymayan doğal görüntü kalitesi değerlerlendiricisi (NIQE) ve görsel karşılaştırma yoluyla filtreler karşılaştırılmış ve en başarılı sonucun olasılıksal pencere tabanlı (PPB) filtrenin verdiği gözlenmiştir.…”

Section: Introductionunclassified

Comparison of despeckle filters for endometrial ultrasound images

Temizkan

Bilge

Temizkan

2015

2015 23nd Signal Processing and Communications Applications Conference (SIU)

View full text Add to dashboard Cite

Özetçe-Endometrium kanseri kadınlarda görülen bir kanser türü olup ultrason görüntüleme yöntemi ile teşhisi yapılabilmektedir. Ultrason görüntüleme, hastaya zararı olmayan medikal teşhis yöntemlerinin başında gelmektedir. Fakat benek gürültüleri, kontrast-gürültü oranını azaltarak ultrason görüntüsünün kalitesini düşürmekte bu da doktorun görüntüyü yorumlamasını etkilemektedir. Bu nedenle benek gürültülerini gidermek ultrason görüntü işlemenin kritik adımlarından biridir. Araştırmacılar tarafından bir çok benek gürültüsü giderme filtresi tasarlanmış fakat endometrium ultrason görüntüleri üzerinde detaylı bir çalışma yapılmamıştır. Bu çalışmada benek gürültüsünü temizlemek için PMAD, SRAD, DPAD ve PPB filtreleri, endometrial ultrason görüntüleri üzerinde denenmiştir. Hangi filtrenin daha etkili çalıştığını belirlemek için NIQE adı verilen gürültüsüz referans görüntüye ihtiyaç duymayan görüntü kalitesi metriği hesaplanmış ve deneysel sonuçlar sunulmuştur. Anahtar Kelimeler -ultrason görüntüsü, benek gürültü, benek gürültü filtreleri, doğal görüntü kalitesi metriği Abstract-Endometrial cancer is a type of cancer in women. Diagnosis of the cancer can be made via ultrasound imaging.Ultrasound imaging is one of the essential methods for noninvasive medical diagnosis. However, speckle noise reduces the quality of the ultrasound image by reducing contrast-to-noise ratio and it also affects image interpretation by doctor. Thus, despeckle is critical phase in pre-processing of ultrasound images. Many speckle noise reduction filters has been designed by the researchers, but it has not been a detailed study on endometrial ultrasound images. In this study, PMAD, SRAD, DPAD and PPB filters are tested on endometrial ultrasound images to reduce speckle noise. NIQE image quality metrics that does not need noise free reference image were computed in order to determine which despeckling algorithm is most effective and experimental results have been demonstrated.

show abstract

Multi-scale deep learning ensemble for segmentation of endometriotic lesions

Podda,

Balia,

Barra

et al. 2024

Neural Comput & Applic

View full text Add to dashboard Cite

Ultrasound is a readily available, non-invasive and low-cost screening for the identification of endometriosis lesions, but its diagnostic specificity strongly depends on the experience of the operator. For this reason, computer-aided diagnosis tools based on Artificial Intelligence techniques can provide significant help to the clinical staff, both in terms of workload reduction and in increasing the overall accuracy of this type of examination and its outcome. However, although these techniques are spreading rapidly in a variety of domains, their application to endometriosis is still very limited. To fill this gap, we propose and evaluate a novel multi-scale ensemble approach for the automatic segmentation of endometriosis lesions from transvaginal ultrasounds. The peculiarity of the method lies in its high discrimination capability, obtained by combining, in a fusion fashion, multiple Convolutional Neural Networks trained on data at different granularity. The experimental validation carried out shows that: (i) the proposed method allows to significantly improve the performance of the individual neural networks, even in the presence of a limited training set; (ii) with a Dice coefficient of 82%, it represents a valid solution to increase the diagnostic efficacy of the ultrasound examination against such a pathology.

show abstract

An automatic segmentation of endometrial cancer on ultrasound images

Cited by 5 publications

References 12 publications

Multi-Scale Deep Learning Ensemble for Segmentation of Endometriotic Lesions

Multi-Scale Deep Learning Ensemble for Segmentation of Endometriotic Lesions

Comparison of despeckle filters for endometrial ultrasound images

Multi-scale deep learning ensemble for segmentation of endometriotic lesions

Contact Info

Product

Resources

About