This document presents a study comparing different graph search algorithms for solving mazes represented as images. The paper implements bidirectional versions of breadth-first search (BFS) and A* search and compares their performance on 8x8 and 16x16 mazes to the traditional unidirectional algorithms. For smaller 8x8 mazes, BFS performed best but for larger 16x16 mazes, bidirectional BFS was most efficient at finding the shortest path. Bidirectional search improves results but uses more space. The key aspect is finding the meeting point where the two searches meet, guaranteeing a solution if one exists.