Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 5 of 5 for “"Fine-Grained Visual Classification"”.
-
Plant Species Identification In The Wild Based On Images Of Organs
… small inter-class image variability makes this a fine-grained visual classification problem. One way to cope with this variability and to reduce image background noise is to predict species based on the plant organs that are visible. In this work, we designed a set of classifiers that predict …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
Fine-grained visual classification in open-set domains remains a persistent challenge for vision systems, especially under limited supervision. In this work, we explore a post-hoc correction framework that combines a strong semi-supervised few-shot discriminative model (SWIFT) with a powerful …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
Fine-grained visual classification in open-set domains remains a persistent challenge for vision systems, especially under limited supervision. In this work, we explore a post-hoc correction framework that combines a strong semi-supervised few-shot discriminative model (SWIFT) with a powerful …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
Fine-grained visual classification in open-set domains remains a persistent challenge for vision systems, especially under limited supervision. In this work, we explore a post-hoc correction framework that combines a strong semi-supervised few-shot discriminative model (SWIFT) with a powerful …
-
Improving Semi Supervised Few-Shot Recognition by Post-Hoc Correction
Fine-grained visual classification in open-set domains remains a persistent challenge for vision systems, especially under limited supervision. In this work, we explore a post-hoc correction framework that combines a strong semi-supervised few-shot discriminative model (SWIFT) with a powerful …