Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference
Zero-shot recognition (ZSR) aims to perform visual classification by category in the absence of training samples. The focus in most traditional ZSR models is using semantic knowledge about familiar categories to represent unfamiliar categories with only the visual appearance of an unseen object. In...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-10-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/10/20/7234 |
| _version_ | 1797550703869165568 |
|---|---|
| author | Doo Soo Chang Gun Hee Cho Yong Suk Choi |
| author_facet | Doo Soo Chang Gun Hee Cho Yong Suk Choi |
| author_sort | Doo Soo Chang |
| collection | DOAJ |
| description | Zero-shot recognition (ZSR) aims to perform visual classification by category in the absence of training samples. The focus in most traditional ZSR models is using semantic knowledge about familiar categories to represent unfamiliar categories with only the visual appearance of an unseen object. In this research, we consider not only visual information but context to enhance the classifier’s cognitive ability in a multi-object scene. We propose a novel method, <i>contextual inference</i>, that uses external resources such as knowledge graphs and semantic embedding spaces to obtain similarity measures between an unseen object and its surrounding objects. Using the intuition that close contexts involve more related associations than distant ones, distance weighting is applied to each piece of surrounding information with a newly defined distance calculation formula. We integrated contextual inference into traditional ZSR models to calibrate their visual predictions, and performed extensive experiments on two different datasets for comparative evaluations. The experimental results demonstrate the effectiveness of our method through significant enhancements in performance. |
| first_indexed | 2024-03-10T15:33:06Z |
| format | Article |
| id | doaj.art-063034f1e094434c829359dce5bbe48e |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T15:33:06Z |
| publishDate | 2020-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-063034f1e094434c829359dce5bbe48e2023-11-20T17:24:17ZengMDPI AGApplied Sciences2076-34172020-10-011020723410.3390/app10207234Zero-Shot Recognition Enhancement by Distance-Weighted Contextual InferenceDoo Soo Chang0Gun Hee Cho1Yong Suk Choi2Artificial Intelligence Laboratory, Hanyang University, Seoul 04763, KoreaArtificial Intelligence Laboratory, Hanyang University, Seoul 04763, KoreaArtificial Intelligence Laboratory, Hanyang University, Seoul 04763, KoreaZero-shot recognition (ZSR) aims to perform visual classification by category in the absence of training samples. The focus in most traditional ZSR models is using semantic knowledge about familiar categories to represent unfamiliar categories with only the visual appearance of an unseen object. In this research, we consider not only visual information but context to enhance the classifier’s cognitive ability in a multi-object scene. We propose a novel method, <i>contextual inference</i>, that uses external resources such as knowledge graphs and semantic embedding spaces to obtain similarity measures between an unseen object and its surrounding objects. Using the intuition that close contexts involve more related associations than distant ones, distance weighting is applied to each piece of surrounding information with a newly defined distance calculation formula. We integrated contextual inference into traditional ZSR models to calibrate their visual predictions, and performed extensive experiments on two different datasets for comparative evaluations. The experimental results demonstrate the effectiveness of our method through significant enhancements in performance.https://www.mdpi.com/2076-3417/10/20/7234zero-shot recognitionsimilarity measuresdistance-weightingknowledge graphsemantic embedding |
| spellingShingle | Doo Soo Chang Gun Hee Cho Yong Suk Choi Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference Applied Sciences zero-shot recognition similarity measures distance-weighting knowledge graph semantic embedding |
| title | Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference |
| title_full | Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference |
| title_fullStr | Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference |
| title_full_unstemmed | Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference |
| title_short | Zero-Shot Recognition Enhancement by Distance-Weighted Contextual Inference |
| title_sort | zero shot recognition enhancement by distance weighted contextual inference |
| topic | zero-shot recognition similarity measures distance-weighting knowledge graph semantic embedding |
| url | https://www.mdpi.com/2076-3417/10/20/7234 |
| work_keys_str_mv | AT doosoochang zeroshotrecognitionenhancementbydistanceweightedcontextualinference AT gunheecho zeroshotrecognitionenhancementbydistanceweightedcontextualinference AT yongsukchoi zeroshotrecognitionenhancementbydistanceweightedcontextualinference |