Skythianos's repositories
FRIQA-ActMapFeat
Full-reference image quality assessment based on convolutional activation maps.
Manipulating-objects-using-keypoints
Finding objects in an arbitrary environment is one of the unsolved problems about robots operating in such environments, e.g. households. In this project a robotics application is presented. The software controlls a robotic arm, and estimates the spatial position and orientation of an object for which it has been trained previously. The estimation is done using images retrieved from a camera mounted on the end effector of the robot. The software uses PnP algorithm which estimates the spatial pose from object points with known 3D coordinates and the corresponding image points. The image points are found via SURF keypoint detector. During training the algorithm, 3D reconstruction is done via multi-view triangulation using multiple images taken from known positions.
Multi-Pooled-Inception-Features-for-No-Reference-Image-Quality-Assessment
Multi-pooled Inception Features for No-reference image quality assessment
DF-CNN-IQA
No-Reference Image Quality Assessment with Convolutional Neural Networks and Decision Fusion
Arteries-based-Traffic-Control-Methods
In this thesis there will be introduced two main traffic control methods, both were designed for networks with arteries. The permeability of these networks is bigger, and also the capability of travelling faster motivates the drivers to use the arteries thanks to the correct traffic light phases. Therefore fewer roads are used intensively and needs to be maintained more often. The first observed algorithm was designed for really specific conditions; therefore the usage of this method is limited. Due to this property in these conditions this method is more efficient, than the others were designed for more general situations. On the oth- er hand the second observed algorithm can be used in general situations without any serious restriction. The first and maybe the most important restriction of the first method is that the traf- fic can flow only in one direction in one artery. In the case of two way traffic, there is a chance, that the side streets will be totally blocked. The main purpose of the method is to serve the arterial traffic in the most efficient way, but it is not acceptable, that side streets do not get green light at all. Other typical property of the algorithm is it is based on prediction, so the algorithm tries to figure out how will look like the traffic in the future based on the traffic right now and some statistics information which is the result of previous measurements. The lights will be controlled based on these predic- tions. The further we want to optimize, the greater will be the inaccuracy. The other method was designed for two way arteries, and the prediction gets less at- tention, because in this case the phases of the lights are determined based on the network, not the traffic. Basically here we need to figure out one cycle of the phases and it will be used for the whole simulation. Therefore this method is less efficient, but here we have the opportunity to use this algorithm for the case, when the network contains intersecting arteries, or a whole arterial loop.
Benford-IQA
Analysis of Benford’s Law for No-Reference Quality Assessment of Natural, Screen-Content, and Synthetic Images
IIRFilter-and-FIRFilter-in-MATLAB-mex
IIRFilter and FIRFilter implemented in MATLAB mex files.
Player-tracking-for-football
Player tracking part of a football analysis software. The player tracking problem is based on the combination of detection and tracking, including preprocessing of the video sequences coming from a fixed camera layout.
Benford-VQA
No-Reference Video Quality Assessment Based on Benford’s Law and Perceptual Features
Calculating-distances-between-sets-of-vectors
It calculates the distance between sets of vectors.
LGV
Full-Reference Image Quality Assessment Based on Grünwald–Letnikov Derivative, Image Gradients, and Visual Saliency
Pixelization
Pixelization is a process that decreases the resolution of an image or video with specific algorythms, thus achieving an image or video containing a smaller amount of pixels. The world of pixelart is a popular topic nowadays. Several image editing applications have built-in modules for creating pixelart from high quality images. In these editors algorythms range from the simplest to the more complex ones.
RGB-Color-Gamut-Visualizer
It creates a 3D scatter plot of an RGB color gamut of a color image.
skythianos.github.io
website
SWLGV
Full-Reference Image Quality Assessment Based on Grünwald–Letnikov Derivative, Image Gradients, and Visual Saliency
Video-handling-in-MATLAB2017b
This MATLAB script reads a video sequnece frame by frame.