Object detection and identification is a fundamental workflow in Computer vision. Proposed system specifies the methods that have been used to overcome the disadvantages that were present in the old methodologies like: e.g R-CNN, KNN etc.
Training is done via convolutional neural network with inception-v3 model with the labels of their respective object. Segmentation is applied as a pre-processing stepwhere the image is divided into equal patches. A rectangle is drawn on the image showing the patches of image data. Image is cropped using the co-ordinates of the rectangle. This method increases accuracy because only the part of an image is used for CNN.