Extracting information from text within images is essential for numerous real-world applications. However, conventional methods often struggle to effectively leverage both text and image to improve accuracy and performance. This research introduces a novel model that overcomes these limitations by integrating graph attention networks, convolutional neural networks, and transformers. The proposed approach constructs detailed graphs from text regions and seamlessly combines image features extracted by convolutional neural networks with text features processed through DistilBERT. Experimental results demonstrate that the proposed model outperforms existing methods, delivering exceptional performance.