Image Detection and Recognition Using Convolutional Neural Networks

The method of collecting useful information after converting an image into digital form is called image processing. This is achieved by performing some operations on the digitized data. Object detection and recognition are the tasks that are important and challenging and has various applications in computer vision. Object detection is verifying the presence of an object in an image and then recognizing the object. Object recognition aims at classifying images and helps in understanding images. Classifying objects and tracking their motion are an important aspect of any intelligent system. This paper proposes a VGG based-convolutional neural network with 12 layers that can be used for image recognition. The proposed architecture is compared with other standard CNNs and is found to be efficient and computationally less complex. Competitive datasets CIFAR-10 and Caltest 101 have been used to validate the model and test the accuracy of the model under different variations. The edges of the images have been enhanced using a Sobel operator before image recognition to improve the efficiency of the architecture.