| Titre : |
Enhancing deep learning based classifiers using out of distribution data detection |
| Type de document : |
document électronique |
| Auteurs : |
Abdelghani Halimi, Auteur ; Ahmed Hadjadj, Auteur ; Sid-Ahmed Berrani, Directeur de thèse |
| Editeur : |
[S.l.] : [s.n.] |
| Année de publication : |
2022 |
| Importance : |
1 fichier PDF (10.3 Mo) |
| Présentation : |
ill. |
| Note générale : |
Mode d'accès : accès au texte intégral par intranet.
Mémoire de Projet de Fin d’Études : Électronique : Alger, École Nationale Polytechnique : 2022
Bibliogr. f. 80 - 84 . - Annexe f. 85 - 94 |
| Langues : |
Anglais (eng) |
| Mots-clés : |
Neural network Out-of-distribution detection Image data Comparative evaluation Classification Proposed method Web interface |
| Index. décimale : |
PN00322 |
| Résumé : |
In this work, three out-of distribution detection methods are implemented, evaluated and compared on several common benchmarks (different natural image datasets), as well as on the ImageNet-O dataset, a novel dataset that has been created to aid research in OOD detection for ImageNet models. In this thesis, we also investigate the effect of label space size on the OOD detection performance, for that we used three different in-distribution datasets (CIFAR-10, CIFAR-100 and ImageNet-1K), and we showed that the performance degrades rapidly as the number of in-distribution classes increases. We concluded by proposing a method that surpasses the three previous methods in detection performances and by creating a web user interface to test out our OOD detection method. |
Enhancing deep learning based classifiers using out of distribution data detection [document électronique] / Abdelghani Halimi, Auteur ; Ahmed Hadjadj, Auteur ; Sid-Ahmed Berrani, Directeur de thèse . - [S.l.] : [s.n.], 2022 . - 1 fichier PDF (10.3 Mo) : ill. Mode d'accès : accès au texte intégral par intranet.
Mémoire de Projet de Fin d’Études : Électronique : Alger, École Nationale Polytechnique : 2022
Bibliogr. f. 80 - 84 . - Annexe f. 85 - 94 Langues : Anglais ( eng)
| Mots-clés : |
Neural network Out-of-distribution detection Image data Comparative evaluation Classification Proposed method Web interface |
| Index. décimale : |
PN00322 |
| Résumé : |
In this work, three out-of distribution detection methods are implemented, evaluated and compared on several common benchmarks (different natural image datasets), as well as on the ImageNet-O dataset, a novel dataset that has been created to aid research in OOD detection for ImageNet models. In this thesis, we also investigate the effect of label space size on the OOD detection performance, for that we used three different in-distribution datasets (CIFAR-10, CIFAR-100 and ImageNet-1K), and we showed that the performance degrades rapidly as the number of in-distribution classes increases. We concluded by proposing a method that surpasses the three previous methods in detection performances and by creating a web user interface to test out our OOD detection method. |
|
PN00100 
Ajouter le résultat dans votre panier
Visionner les documents numériques
Faire une suggestion Affiner la rechercheEnhancing deep learning based classifiers using out of distribution data detection (2022)
