Tools
Tools
Pataskar, Utsav (2024) Depth Estimation for indoor environments using Augmented and Regularized Data through Knowledge Distillation. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (9MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (1MB) | Preview |
Abstract
Depth Estimation is one of the important applications of computer vision which are further used in autonomous vehicles, robotics vision and AR/VR world. This research focuses on increasing generalization capabilities of depth estimation models on indoor settings which have low lightings, clustered and occluded objects and overall lack the diversity in terms of texture, has consistent and repetitive structural geometry. We deployed teacher-student framework to implement a ResNet-based pre-trained model as the teacher which will generate it’s own pseudo depth maps from NYU-Depth V2 and Augmentations. The student model DenseDepth-169 based on U-Net learns from the teacher model and it’s predictions. The proposal addresses overfitting and generalization problem by employing data augmentation and dropout regularization and increasing overall dataset size significantly. Edge Detection and contrast adjustment further aid in improving input feed quality. The research also provides a base for scalable and efficient indoors depth estimation models that are adaptive to diverse environments.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Raj, Kislay UNSPECIFIED |
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence > Computer vision Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence > Computer vision |
| Divisions: | School of Computing > Master of Science in Artificial Intelligence |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 20 Jun 2025 09:16 |
| Last Modified: | 20 Jun 2025 09:16 |
| URI: | https://norma.ncirl.ie/id/eprint/7958 |
Actions (login required)
 |
View Item |