Recognition and Recombination of Graphic Design Elements Using Computer Vision.

Computer vision in artificial intelligence (AI) is concerned with teaching computers to see and make sense of their visual surroundings. Machines can recognize, understand, and reconstruct objects and actions with the help of computer vision systems that process digital images captured by cameras and other visual inputs. Graphic design elements are essential to produce visually appealing and functional designs. Recognizing and skillfully applying these components is critical for any graphic designer. This work aimed to acknowledge and recombine visual design elements using computer vision. First, we examined the imaging data of Egyptian bananas gathered from the Fruit Research and Extension Center at Penn State University. We used histogram equalization for images to enhance the processing, which improved local contrast and structure visualization. We use the oust threshold approach to detect image data divided with separate values by images after splitting an image into many segments for easy analysis. We analyzed images using the gray-level co-occurrence matrix (GLCM) for recombinant three-dimensional (3D) shapes and structures from 2D images and principal component analysis (PCA) for recognition from a digital image in feature extraction. We suggested the hybrid dove swarm red fox optimization with convolutional neural network (HDSRFO-CNN) to use computer vision to identify and repurpose graphic design elements. When compared to existing methods, the suggested performed well. This allows for the evaluation of the suggested solution's accuracy, precision, specificity, sensitivity, and AUC-ROC curve. The proposed model had a 98% accuracy rate. This implies that compared to the existing methods, which include 72% of CNN-LSTM-GA, 80% of DCNN, and 89% of ANN-GAN with more images. [ABSTRACT FROM AUTHOR]