Image Segmentation using Convolutional Neural Network for Image Annotation
Sandip Nemade, Shefali Pratap Sonavane · 2019
Due to rapid advancement in the digital display, communication and storage devices with effective techniques are needed to organize, index, retrieve and annotate a large image database. Image segmentation finds application in various areas of image processing and computer vision. Inferring of low level features from the given image is a challenging task in unstructured regions. Most of the annotation and retrieval algorithms fail to consider region semantics. This paper proposes convolutional neural network (CNN) based image segmentation for image annotation application. The proposed CNN includes pixel based prediction of the regions that are applied to obtain low level image features. The algorithm uses image region information based on the precise color distribution within the image. Experimental results demonstrate better results of image segmentation using CNN.