Historical OCR engines have their accuracy lying between 70-80% for a high-quality image at page level. That means in a page of 100 words 70-80 words are accurate. This will lead to significant inaccuracies if used on a large volume of sensitive documents. The project aims to improve the accuracy of the result by implementing an ideal combination of image preprocessing techniques prior to text extraction. The aim is to extend existing pipelines to improve the accuracy by; improving the quality of the image and using Natural Language Processing to extract information from raw text.

Before the development of this project, there were many loopholes in the process of taking attendance using the traditional methods which caused troubles to most of the institutions. Automated Attendance System using MTCNN and FaceNet is expected to be able to replace the old manual attendance process, which is currently used. The system has also proved to be time saving, securing and conquering the defects by merely saving resources but also reducing human intervention in the whole process. This standalone system detects the person which was already given in the dataset to track and an embedding being created was successfully detected with an accuracy of 91.304%. In real time scenarios, MTCNN and FaceNet algorithms outperforms other algorithms with better recognition rate and low false positive rate. SVM has also proved to be a better classifier when compared to others.

The current university election system lacks proper voter verification, allowing anyone with a university ID to vote, which can lead to ineligible voting. Limited interaction between candidates and voters through infrequent rallies leaves many students uninformed about their choices. Additionally, manual vote counting is time-consuming, and there is potential for vote manipulation by authorities. 

The proposed online system aims to improve this by introducing secure candidate registration and login, enhancing voter-candidate interaction through an interactive platform, and providing real-time tallying and graphical representation of election results along with its sentiment analysis. This system will ensure a fairer, more efficient, and transparent election process.