Auto-Tagging of Images using Machine Learning
The proposed system uses Machine Learning algorithms to make the machine capable of learning on its own to categorize images according to the object/scene in the image. This software system will be used to automatically categorize images based on the object/scene in the image. This system can be used in cases where large-scale image datasets need to be classified accordingly. By sufficiently training the system with sample datasets over time, the accuracy of the software can be improved with use.
The software uses concepts of Machine Learning to learn on its own, thereby reducing the amount of human intervention required. As a result, the accuracy of the software will be improved as more datasets are given to the software. The system will first convert the image data to relevant feature vectors that an ML framework works on, to generate an output that can classify the image at hand accordingly. Thus, the system can be used to automatically tag images after sufficient training.
This system uses the scikit-learn library for ML and the scikit-image library for Image Processing. The pexif library is used to write tags to the image file
This was done as a Mini-Project for Computer Science and Engineering at College of Engineering Chengannur, by Bhargav MD, Bony Kuriakose, Lakshmi S and Leroy J Vargis