An Enhanced HMAX Model to Improve Object Recognition

A new approach is proposed within this paper to accelerate the process of a biologically inspired Hieratical Model and X (HMAX), which is a biologically stimulated method, this model has been an efficient procedure for object recognition. Object recognition is beneficial in numerous applications such as driverless cars, and security systems. However, patches in HMAX are extracted randomly, causing in production redundant and uninformed extracted patches. Also, there is not any mechanism to evaluate them, these play a big challenge step for achieving high recognition accuracy. This paper presents a technique for selecting specific kinds of features of an image established on a Shi-Tomasi Corner Detector. A proposed approach is applied to the Caltech5 dataset. In the training stage, 70% of images are used. The results appear significant achievements in object recognition accuracies.