Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation
Colorectal cancer is one of the leading cancer death causes worldwide, but its early diagnosis highly improves the survival rates. The success of deep learning has also benefited this clinical field. When training a deep learning model, it is optimized based on the selected loss function. In this wo...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-08-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/8/8/1316 |
| _version_ | 1797559624439693312 |
|---|---|
| author | Luisa F. Sánchez-Peralta Artzai Picón Juan Antonio Antequera-Barroso Juan Francisco Ortega-Morán Francisco M. Sánchez-Margallo J. Blas Pagador |
| author_facet | Luisa F. Sánchez-Peralta Artzai Picón Juan Antonio Antequera-Barroso Juan Francisco Ortega-Morán Francisco M. Sánchez-Margallo J. Blas Pagador |
| author_sort | Luisa F. Sánchez-Peralta |
| collection | DOAJ |
| description | Colorectal cancer is one of the leading cancer death causes worldwide, but its early diagnosis highly improves the survival rates. The success of deep learning has also benefited this clinical field. When training a deep learning model, it is optimized based on the selected loss function. In this work, we consider two networks (U-Net and LinkNet) and two backbones (VGG-16 and Densnet121). We analyzed the influence of seven loss functions and used a principal component analysis (PCA) to determine whether the PCA-based decomposition allows for the defining of the coefficients of a non-redundant primal loss function that can outperform the individual loss functions and different linear combinations. The eigenloss is defined as a linear combination of the individual losses using the elements of the eigenvector as coefficients. Empirical results show that the proposed eigenloss improves the general performance of individual loss functions and outperforms other linear combinations when Linknet is used, showing potential for its application in polyp segmentation problems. |
| first_indexed | 2024-03-10T17:47:57Z |
| format | Article |
| id | doaj.art-7c08f93cb7af41a3b2c9e6b104bdd341 |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-10T17:47:57Z |
| publishDate | 2020-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-7c08f93cb7af41a3b2c9e6b104bdd3412023-11-20T09:27:46ZengMDPI AGMathematics2227-73902020-08-0188131610.3390/math8081316Eigenloss: Combined PCA-Based Loss Function for Polyp SegmentationLuisa F. Sánchez-Peralta0Artzai Picón1Juan Antonio Antequera-Barroso2Juan Francisco Ortega-Morán3Francisco M. Sánchez-Margallo4J. Blas Pagador5Jesús Usón Minimally Invasive Surgery Centre, N-521, km 41.7, E-10071 Cáceres, SpainTECNALIA, Basque Research and Technology Alliance (BRTA), Parque Tecnológico de Bizkaia, C/Geldo. Edificio 700, E-48160 Derio, SpainDidactics of Mathematics, University of Cadiz, Avda. República Saharaui s/n. Campus de Puerto Real, E-11519 Puerto Real, SpainJesús Usón Minimally Invasive Surgery Centre, N-521, km 41.7, E-10071 Cáceres, SpainJesús Usón Minimally Invasive Surgery Centre, N-521, km 41.7, E-10071 Cáceres, SpainJesús Usón Minimally Invasive Surgery Centre, N-521, km 41.7, E-10071 Cáceres, SpainColorectal cancer is one of the leading cancer death causes worldwide, but its early diagnosis highly improves the survival rates. The success of deep learning has also benefited this clinical field. When training a deep learning model, it is optimized based on the selected loss function. In this work, we consider two networks (U-Net and LinkNet) and two backbones (VGG-16 and Densnet121). We analyzed the influence of seven loss functions and used a principal component analysis (PCA) to determine whether the PCA-based decomposition allows for the defining of the coefficients of a non-redundant primal loss function that can outperform the individual loss functions and different linear combinations. The eigenloss is defined as a linear combination of the individual losses using the elements of the eigenvector as coefficients. Empirical results show that the proposed eigenloss improves the general performance of individual loss functions and outperforms other linear combinations when Linknet is used, showing potential for its application in polyp segmentation problems.https://www.mdpi.com/2227-7390/8/8/1316deep learningloss functionsprincipal component analysispolyp segmentation |
| spellingShingle | Luisa F. Sánchez-Peralta Artzai Picón Juan Antonio Antequera-Barroso Juan Francisco Ortega-Morán Francisco M. Sánchez-Margallo J. Blas Pagador Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation Mathematics deep learning loss functions principal component analysis polyp segmentation |
| title | Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation |
| title_full | Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation |
| title_fullStr | Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation |
| title_full_unstemmed | Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation |
| title_short | Eigenloss: Combined PCA-Based Loss Function for Polyp Segmentation |
| title_sort | eigenloss combined pca based loss function for polyp segmentation |
| topic | deep learning loss functions principal component analysis polyp segmentation |
| url | https://www.mdpi.com/2227-7390/8/8/1316 |
| work_keys_str_mv | AT luisafsanchezperalta eigenlosscombinedpcabasedlossfunctionforpolypsegmentation AT artzaipicon eigenlosscombinedpcabasedlossfunctionforpolypsegmentation AT juanantonioantequerabarroso eigenlosscombinedpcabasedlossfunctionforpolypsegmentation AT juanfranciscoortegamoran eigenlosscombinedpcabasedlossfunctionforpolypsegmentation AT franciscomsanchezmargallo eigenlosscombinedpcabasedlossfunctionforpolypsegmentation AT jblaspagador eigenlosscombinedpcabasedlossfunctionforpolypsegmentation |