Leveraging CNN for Accurate Hand Gesture Recognition

Human-computer interaction systems rely mostly on hand gesture recognition since they give people a natural and simple way to communicate with digital devices. In this work, a Convolutional Neural Network (CNN) based model is developed to exactly classify hand gestures from a dataset. With an accuracy, precision, recall, and F1-score of 1.0 on the test set, the model achieved a perfect classification performance by means of deep learning approaches, so confirming its capacity to precisely detect every motion free from mistakes. The end loss value of 1.33e-09 validates the ability of the model to lower classification error. By means of data augmentation techniques, the durability of the model is enhanced and overfitting is prevented, therefore ensuring generalisation over a spectrum of inputs. Further, our CNN model is compared with standard machine learning methods including Support Vector Machines (SVM) and K-Nearest Neighbours (KNN). It is observed that the proposed method outperforms the existing methods with improved feature extraction capacity. In future, the focus will be on broadening the dataset for real-time detection and recognition of dynamic gestures in progressively challenging environments.