High performance of optimizers in deep learning for cloth patterns detection

In Deep learning, optimization methods are an essential role. Optimizers used to change weights and learn rates reduce or minimize losses in a neural network. Nowadays, Learning is widely used, especially object detection tasks. this case, cloth patterns considered for assist visually impaired people. The person had many limitations when doing their activity, not least choosing clothes; it would be difficult without guidance tools like Braille labels. research, system was researched detect 11 different (Argyle, Batik, Camouflage, Gingham, Dotted, Floral, Leopard, Solid, Striped, Zebra, Zigzag) using RetinaNet with ResNet-152 architecture. To achieve the best model performance, compare 6 optimizers, such as Stochastic Gradient Descent, Root Mean Square Propagation, Adaptive Moment Estimation, Delta, Norm, conducted. Each optimizer has trained three learning (1E-3, 1E-4, 1E-5). A Adamax rate 1E-4 achieved highest accuracy mAP (mean Average Precision) 91.28% during training process. Based on testing result precision 93.01%, recall 92.91%, F1-Score 92.79%, 92.91%.