Nowadays what all we witness is based on detection of objects, either still or mobile. The information gathered from the detection is further scrutinized for further studies. This detection process is purely a wonder of technology known as Machine learning and its advance version known as Deep learning. Detection is usually used for surveillance, robotics, image retrieval, driver assistance and object tracking in real-time world. Detection is bifurcated in two categories i.e. localization and classification. An image may contain multiple objects for which localization is needed. Localization is done in an optimized way such that it localizes most of the objects seen in the image. In this method we will be using Tensor Flow and Coco dataset that will be labelling thousands of the localized objects according to their respective class with accuracy being checked at the same time. It may result in some difficulties such as overlapping of multiple objects while localization, misinterpretation of objects that may result in incorrect labelling, or grouping of objects will lead to overlooking of some objects. In this venture we will create services that would be more preferable in terms of detection and accuracy.