Grip-Vision: Enhancing Object Handling with Computer Vision-Based Grip Classification

The hand is the primary functional part of the human body. Due to various traumatic injuries or accidents, people are suffering from upper extremity amputations. This paper presents a detailed description of the system which is to be integrated with the artificial hand for amputees. This computer vision-based approach is aimed at classifying objects based on the grip required for lifting them using the canny edge detection method and feature descriptors like BRISK. Experimentation is done on household objects like bottles, cups, and mobile phones with the help of state of art Machine Learning models and classifiers like Random Forest, SVM, Decision Tree, and KNN. The trained model can classify the objects based on their required grips with the highest accuracy scores of both the KNN classifier and Random Forest classifier that is 91.86% and 90.27% respectively with the proper output of the classified grip.