On the Generalization Capabilities of FSL Methods Through Domain Adaptation: A Case Study in Endoscopic Kidney Stone Image Classification

Mendez-Ruiz, Mauricio; Lopez-Tiro, Francisco; Flores-Araiza, Daniel; Jonathan, El-Beze,; Ochoa-Ruiz, Gilberto; González-Mendoza, Miguel; Hubert, Jacques; Méndez-Vázquez, Andres; Daul, Christian

doi:10.1007/978-3-031-19493-1_21

Cited by 5 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As noticeable in Table 9, the proposed DL-based models obtained outstanding results: a general performance of 95 ± 08%, 94 ± 08%, and 97 ± 03% was respectively reached for surface, section, and mixed views of in-vivo images of four classes. The proposed DL-model led to a higher precision than most of the models that classify kidney stones using exvivo images (average precision of 74% for up to 5 classes in [14] and average precision of 84% for five classes in Black et al [15], see Table 3) The proposed model also outperformed other methods tested on in-vivo images (average precision of 88% for up to 6 classes in [18]). The work in [19] achieved a high precision of 88%, 84%, and 96% for surface, section, and mixed views of six classes.…”

Section: Discussionmentioning

confidence: 83%

“…Fragmentation and dusting can be performed with the same laser, the difference being in the settings of the instrument. In the dusting mode, low energy (0.2-0.5 J) and high frequency (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) are used to vaporize the kidney stones [35].…”

Section: A Context and Recent Trends In Ureteroscopymentioning

confidence: 99%

“…In a more recent work [18], a novel metric learning approach was proposed for the classification of endoscopic kidney stone images. The authors proposed a method to generalize the classification of kidney stones in two different domains.…”

Section: B Previous Attempts Of Kidney Stone Classificationmentioning

confidence: 99%

See 2 more Smart Citations

On the In Vivo Recognition of Kidney Stones Using Machine Learning

Lopez-Tiro,

Estrade,

Hubert

et al. 2024

IEEE Access

Self Cite

View full text Add to dashboard Cite

Determining the type of kidney stones allows urologists to prescribe a treatment to avoid the recurrence of renal lithiasis. An automated in-vivo image-based classification method would be an important step towards an immediate identification of the kidney stone type required as a first phase of the diagnosis. In the literature, it was shown on ex-vivo data (i.e., in very controlled scene and image acquisition conditions) that an automated kidney stone classification is indeed feasible. This pilot study compares the kidney stone recognition performances of six shallow machine learning methods and three deep-learning architectures which were tested with in-vivo images of the four most frequent urinary calculus types acquired with an endoscope during standard ureteroscopies. This contribution details the construction of an in-vivo dataset of endoscopic images with four of the most recurrent classes in clinical practice. It also describes the design, implementation, and results of the classifiers (shallow machine learning and deep learning-based methods) of kidney stones. Even if the best results were obtained by the InceptionV3 architecture (weighted accuracy, precision, recall and F1-Score of 97 ± 03%, 97 ± 03%, 98 ± 04%, and 98 ± 03%, respectively), it is also shown that choosing an appropriate colour space and texture features allows a shallow machine learning method to approach closely the performances of the most promising deep-learning methods (the XGBoost classifier led to weighted accuracy, precision, recall and F1-Score values of 96 ± 14%).

show abstract

Section: Discussionmentioning

confidence: 83%

Section: A Context and Recent Trends In Ureteroscopymentioning

confidence: 99%

See 1 more Smart Citation

On the In Vivo Recognition of Kidney Stones Using Machine Learning

Lopez-Tiro,

Estrade,

Hubert

et al. 2024

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…One of the most prevalent health issues is kidney stone disease 1,2 . This rate has been observed to range amid 1 and 20% in prevalence studies 3 .…”

Section: Introductionmentioning

confidence: 99%

Self‐attention based progressive generative adversarial network optimized with arithmetic optimization algorithm for kidney stone detection

V.¹,

K.²,

Rooban

et al. 2023

Concurrency and Computation

View full text Add to dashboard Cite

Summary Self‐attention based progressive generative adversarial network optimized with arithmetic optimization algorithm (AOA) is proposed in this manuscript for kidney stone detection. Initially, the input kidney stone images are gathered via CT kidney dataset. Then, the input image is preprocessed by utilizing APPDRC filtering approach. Also, the preprocessed images are given to the multi‐level thresholding segmentation technique for segmented the image. Then the segmented images are given to the TF‐IDF feature extraction method for extracting the features. Then the extracted features are fed to feature selection using Weibull distributive generalized multidimensional scaling methods for selecting the features. Then the selected features are SPGGAN classification method for kidney stone detection. Generally, SPGGAN not reveal any adoption of optimization methods compute optimum parameters for assuring correct kidney stone detection. Thus, AOA is used for optimizing the weight parameters of SPGGAN and it is implemented in python, its performance is examined under certain performances metrics, such as accuracy, precision, sensitivity, specificity, F‐measure, computational time, and ROC. The proposed SPGGAN‐AOA‐KSD method attains 54.78%, 34.89%, and 20.96% higher accuracy and 3.45%, 4.08%, and 5.06% greater AUC compared with existing methods, such as a ANN‐OGGA‐KSD, HMANN‐BPA‐KSD, and ANN‐CSOA‐KSD respectively.

show abstract