Abstract: Machine learning and deep learning methods can be applied to various purposes continuously. Present-day methodology for order and ID of objects puts a critical occupation in the field of online business, perception, and various others. One of the fundamental objections of the splendid order and recognizing confirmation is to work on the technique in the high-level world and make human life more directly similar to pleasing and exact in finding the things. The machine learning and deep learning methods used here assist with characterizing characterization according to the predefined classes and progressively recognizing evidence of objects. We reveal the Bag of Features procedure used to find picture depiction. Class prediction precision of contrasting classifiers algorithms is assessed on Caltech images. To incorporate extraction limits, we evaluate the usage of the conventional Speed Up Robust Features system against overall concealing feature extraction