Accuracy in Object Detection by Using Deep Learning Method

Accuracy and Precision measurement of an object in a complex environment for mobile robot applications is an important task. In a complex environment it is difficult to identify similar objects, such as chairs, tables, furniture and other objects available. Different methods are available to identify such objects, but high accuracy and precision are not easily achievable. This paper presents MobileNet with Vision Transformer for identification of objects in complex environments. Several experiments have been conducted for indoor environments where 94.3% accuracy is achieved for indoor environment with the implementation of MobileNet with Vision Transformer model.