Συντάχθηκε 14-07-2014 09:01
από Vasiliki Grigoraki
Email συντάκτη: vgrigoraki<στο>tuc.gr
Ενημερώθηκε:
-
Κύρια: υπάλληλος ΗΜΜΥ.
Άλλες ιδιότητες: Unknown -#-@ΗΜΜΥ
ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ
Σχολή Ηλεκτρονικών Μηχανικών και Μηχανικών Υπολογιστών
Πρόγραμμα Προπτυχιακών Σπουδών
ΠΑΡΟΥΣΙΑΣΗ ΔΙΠΛΩΜΑΤΙΚΗΣ ΕΡΓΑΣΙΑΣ
ΑΝΑΣΤΑΣΟΠΟΥΛΟΥ ΦΟΙΒΟΥ
με θέμα
Implementation of WebP Algorithm on FPGA
Δευτέρα 14 Ιουλίου 2014, 13:00 μ.μ.
Αίθουσα 137Π39, Κτίριο Επιστημών, Πολυτεχνειούπολη
Εξεταστική Επιτροπή
Αναπληρωτής Καθηγητής Παπαευσταθίου Ιωάννης (επιβλέπων)
Καθηγητής Πνευματικάτος Διονύσιος
Καθηγητής Ζερβάκης Μιχαήλ
Abstract
Image processing has become a crucial part of current technology status quo, being used in many fields such as computer vision, computer graphics and other . In modern sciences and technologies, images also gain much broader scopes due to the ever growing importance of scientific visualization . Complex scenarios of data processing in a wide variety of scientific fields indicate the necessity to visualize large data structures in a most effective way.
Considering the massive growth of internet , both as a scientific and industrial field, image processing is critical to the efforts of software developers as great part of information provided in the multimedia based computing industry contains some kind of image form. Think about all the images , videos, etc a person is brought upon in an average internet session every day, and you can estimate the vitality of image processing algorithms in the effort to make web faster and increase our productivity.
In this thesis WebP was studied , a new image format that provides lossless and lossy compression for images on the web , aiming to increase its performance. Focusing on the more frequently used lossy compression that VP8 encoder facilitates, the algorithm's hotspots were analyzed and the critical functions were implemented in hardware using Xilinx's Vivado HLS( High Level Synthesis) . This tool's convenience of use compared to manual hardware design made it possible to quickly implement various architectures and designs for the hardware accelerated modules and then compare them to choose the most optimal solution. Finally, design space exploration was performed to evaluate the resources used and assess the system’s performance.
