Scholarly open access journals, Peer-reviewed, and Refereed Journals, Impact factor 8.14 (Calculate by google scholar and Semantic Scholar | AI-Powered Research Tool) , Multidisciplinary, Monthly, Indexing in all major database & Metadata, Citation Generator, Digital Object Identifier(DOI)
We now live in an era of Artificial Intelligence and the Internet of Things, and our devices generate a huge amount of data. Most of this data is in the form of photos, videos, and images. Most of these image data obtained from cameras and censors are unstructured. Hence, we must depend on lots of modern techniques of machine learning and artificial intelligence to analyze these data and make efficient use of them. Image classification comes as an effective method to analyze and interpret these unstructured image data. Based on some specific rules image classification tries to categorize by assigning labels to groups or pixels or vectors present in an image. Its application can be in various fields such as medical imaging, satellite imagery, traffic control systems, computer vision, and many more.
Recently it has also become a very useful technology for remote sensing. Earlier the resolution of the images obtained from satellites would be very low, and the pixels would not be clear hence it would be difficult to analyze them. Hence researchers also thought of Object-based analysis which broke down the image into meaningful components from the scene that distinguishes the image from others. The scientists discovered that object-based analysis is better compared to the previous method of using pixels. With the advent of machine learning even semantic methods where semantic level classes were defined for the images obtained from remote send=sing and segregated into airplanes, forests, grassland, water bodies, etc.
Deep learning methods made the process faster and simpler as it helps in analyzing and interpreting large amounts of data easily. Its multiple processing layers help to understand more features from the data along with maintaining high-level abstraction. A convoluted Neural Network is a type of deep learning method that can assign various importance to aspects of the image which are also objects and differentiate one from the other. CNN requires lesser preprocessing compared to all other classification models. It is analogous to the neural network of the human brain. In this research, we have trained the CNN model using the train and validation dataset and once it is trained, we use it to predict the images under the prediction folder having images of classes buildings, forests, glaciers, mountains, sea, and street.
Keywords:
image processing, keras, deep learning, CNN
Cite Article:
"Image Classification using Keras", International Journal of Science & Engineering Development Research (www.ijrti.org), ISSN:2455-2631, Vol.7, Issue 10, page no.155 - 162, October-2022, Available :http://www.ijrti.org/papers/IJRTI2210021.pdf
Downloads:
000202613
ISSN:
2456-3315 | IMPACT FACTOR: 8.14 Calculated By Google Scholar| ESTD YEAR: 2016
An International Scholarly Open Access Journal, Peer-Reviewed, Refereed Journal Impact Factor 8.14 Calculate by Google Scholar and Semantic Scholar | AI-Powered Research Tool, Multidisciplinary, Monthly, Multilanguage Journal Indexing in All Major Database & Metadata, Citation Generator
Facebook Twitter Instagram LinkedIn