| dc.description.abstract | In this thesis, we present an automated solution for assembling jigsaw puzzles
using a combination of image processing techniques. Our approach involves
employing feature extraction methods to isolate and categorize the edges of
puzzle pieces. This enables us to analyze and combine the shape, length, and
colour of the edges to determine the correct adjacency relationship between
the pieces. We present the assembled puzzle solution visually to the end user.
To determine the most suitable fit, we utilize a similarity score function
to compare the edges. This score function facilitates the local matching
procedure, allowing us to identify the most likely connections between puzzle
pieces.
In addition to the local assembly method, we also incorporated a global
assembly method. The global assembly approach aimed to optimize the
overall puzzle assembly process and offered a more robust and precise solution
for automated puzzle assembly. This integration of local and global assembly
methods resulted in a reliable puzzle solver for our data set. | |