Document Classification approach is state-of-the-art technique using deep convolutional neural networks and transfer learning. We used weights learned on ImageNet data using framework (VGG-16,19). Handwritten Word Recognition (HWR) consists of convolutional Neural Network (CNN) layers, recurrent Neural Network (RNN) layers and a final Connectionist Temporal Classification (CTC) layer

The project is about categorize complaint text by bank customers in batches. Bank colleague are logging complaint in correct categories. Engine shall able to classify text in proper category so that bank will get to know in which area customer have complaints. Both unsupervised and supervised approach is used to categorize and classifying the complaints.

The project is about predict customer who is going to complaint in coming 7 days based on his transactions and communications from bank.

The project is about detection of human in restricted area in port site. Person is detected using machine vision based approach with thermal camera. System is working in real time. Project included many phases – Sensor selection, data collection, analysis, annotation, algorithm development (pre-processing, segmentation, machine learning classification and tracking), performance evaluation and system integration.

The project is about constrained optimization of one of the process in mining industry to generate maximum revenue. Challenging part is to understand physics involved in the process. Implementation of mathematical model and finance model involved.

There are many types of documents (like Invoices) from different vendors with differnet template. Requirement is to extract the attibutes from this documents.