108 research outputs found
Visual Genome - Visual Relationship Detection - Scene Graph Generation using Message Passing Neural Networks and Graph Convolutional Networks
This repository contains a processed version of Visual Genome for Visual Relationship Detection, from the Diploma (MSc) thesis Scene Graph Generation using Message Passing Neural Networks and Graph Convolutional Networks by Miltiadis Kofinas, supervised by Christos Diou and Anastasios Delopoulos.
The original thesis is written in Greek
Νευρωνικά Δίκτυα Ανταλλαγής Μηνυμάτων και Συνελικτικά Δίκτυα Γράφων για Εξαγωγή Γράφου Σκηνής Εικόνων
Μιλτιάδης Κοφινάς
https://ikee.lib.auth.gr/record/300900
A summarized English version of the thesis can be accessed here.
It contains region proposals for VGG-16 for all images, and metadata about the bounding box distribution and the predicate classes
Optimal Subband Analysis Filters Compensating for Quantization And Additive Noise
In this paper, we present an analysis filter design technique which optimally defines the proper decimator so that the quantization noise is compensated. The analysis is based on a distortion criterion minimization using the Lagrange multipliers. The optimal decimation filters are derived through a Ricatti solution which involves both the quantization and the interpolation filters. Experimental results are presented indicating the good performance of the proposed technique versus conventional subband filter banks in the presence of quantization noise
- …
