Hello everyone. this is Mohi with 17 years of experience in IT industry. today I'm gonna talk about OCR technology and how it is considered as the cornerstone of modern intelligent automation. let's start with what is O-C-R-O-C-R, also known as optical character recognition. Is a technology that converts images of texts, like scan documents or photos into machine learning, machine readable editable text. What it does it allow you to search, copy and edit text from images, making them more, readable, accessible, and usable. OCR enables intelligent automation by digitizing and processing document based information. OCR was valued as at a massive 9.3 billion in 2022 with strong growth expected by 2030 due to increased demand of automation and digital transformation. So that, coming to the topics that I'll be covering today, I'll be covering today the technical foundation, which, in turns cover how the OCR system actually work. we'll also be talking about the framework of using OCR in different industries, and the practical industry application in, healthcare, finance, legal, and also we'll talk about education sector as well. finally we'll discuss how OCR enhances efficiency and digital transformation across these various sectors. coming to the slide, two. this shows, the, graph shows additional adoption, rate across various, continents. as it shows, the North America has the highest adoption rate, of 82%, followed by Europe, which is 78, and then followed by Asia Pacific around 65, person, which is the most emerging potential, that OCR has. news research has shown that AI and machine learning integrate, integration boost OCR growth, and it's gonna be, gonna do it in the coming years as well. deep learning based OCR solution shows high potential. and it's also being discovered that banking, and financial service, which is BFSI domain, has the largest market share in 2022. coming to, how this OCR technical architecture, works and what are the stages involved in it? coming to the first stage, which is known as image pre-processing, this stage also, can be subdivided into further sub stages. let's talk about that first. the first, sub stages ization. Which convert the gray scale image to binary using adaptive thresholding to handle the, lighting variations. The second sub stage is orientation correction, which detects and, correct the SKUs using, HOF transformation and, projection profile analysis. The third sub stage is noise reduction, which is a very important stage. in pre-processing. It removes artifacts using, median filtering. Gaussian smoothing and, more follicle operation. I've used some heavy terms here. I understand that. so let's, take a small deep dive in into it. What is like gian smoothing? it's a technique, that blurs the images or data. By applying, Gaussian function effectively, it reduces the noise and detail while preserving the overall structure. I also talked about, more follicle operation. What exactly it is? It is again, a set of, the non-linear filters, that analyze and modify shapes with operation, including erosion dilation, opening and closing, coming to the fourth, substrate, which is a resolution enhancement, which is self-explanatory. That is, it improves a low resolution images using super resolution and sampling and interpolation. Now, what is the impact of pre-processing? Alright, so it enhances the OCR accuracy by up, by almost like 35%. It also reduces the error rate by 40 to 60%. Especially for the degraded or low quality document, and that is why it's a very important stage of, OCR coming, to the second, stage, which is feature, extraction and recognition. coming to this, what it does, it, it just converts the, raw image into the recognizable features. It uses, again, uses, geometrical features like shapes, angles, strokes, and, topological features like whole endpoint junctions, to effectively, scan the document, right? it also uses some complex scripts and, what it also, the second, stage of this, is character segmentation. What it does is it identifies the characters using connected component analysis and clustering algorithm. yeah, let me just, go through it. what it means. clustering, clustering algorithm, it is again a method, used in unsupervised, machine learning to group, similar, data points into clusters or groups, based on their characteristics without any prior knowledge of the group. and also it also improves handling of vertical and horizontal text coming to pattern recognition. it uses some, some, again, some algorithms, template matching and adaptive algorithms to do that. it, neural networks also improves, recognition across various character sets. And multiple classifier approach approaches enhance, accuracy, coming to the third, and the, the stage which is post, processing. again, a very important stage of OCR. it represents a final critical refinement, layer in modern OCR. Again, it refines raw O-C-R-O-T, OCR output into high confidence and accurate text. what are the techniques which are used? It uses contextual understanding and probability stick analysis to correct the errors and what's the impact of it? it, it is used to, make the document or the OSHA technology. Accurate, which is almost like 95.8% accuracy. It reduces the errors by almost 82%. And it outperforms image pre-processing, which is 92.5% in feature extraction, 88 point, 7%. key role of, post-processing is, it elevates OCR from functional to exceptional, performance. now since we've gone through all the stages of OCR, let's see, what's a practical, implementation, which is used in various industries? let's first talk about, healthcare, which is my favorite one. how is used, what it is used for? the very first thing that it's used for is medical, record management, which, where it reduces the manual entry errors. It digitizes the patient records, prescription insurance form and whatnot. another thing is it provides, clinical decision support. It helps the doctors and access the historical patient, data very easily. It also provides the cr, it's very crucial for the emergency care to provide very quick treatment, to the patients. the third implementation, use is the laboratory result management, which enables, fast and accurate, access to the test results. It ensures, HIPAA compliance and reduces delays. We all are aware how much, importance, HIPAA compliant has in the healthcare sector. so it, it takes care of that. It improves, the patient outcomes as well. So this is all about the healthcare sector, coming to the financial, sector. let's first see the key, highlights about it. we have seen that in financial sector, what it does, it, it provides, 92.7% processing efficiency. what I mean by that? It is, it paper-based, paper-based, processes are not, converted to digital, reducing the bottleneck and resource use. It also gives 99.1%, document accuracy. so OCR technology ensures nearly perfect accuracy in this case, in document processing, reducing errors, compliant risk, and whatnot. it also, has, 76.8%, time reduction. processing times have drastically dis, decreased, enabling the same day service for transaction. That one stop days. third key highlight is 90.3% user satisfaction. financial professionals, report high satisfaction with the OCR system, praising improved workflow and less, administrative work. key impact of OCR in financial sector. coming to that. from the audit perspective, what it does for them is OCR reduces the audit preparation time, up to 70, 70%, from the client. compliance perspective, intelligence system flags, regulatory issues in real time. in banking, it's been seen that oia speed up the customer identification verification. cutting. it also cut down the onboarding times and improves our fraud detection and customer experience as well. coming to the third, sector, which is a lock, legal document management. so what it does is OCR does, document digitization, which is OCR technology, speed up their digitization, and analysis of legal document like contracts. code filing and the case papers. It also helps in a litigation support, which is a good, very good, nice feature, which OCR enables a faster document, discovery by allowing, quick searches and analysis of, large document set, cutting down again on the manual effort. It, it also performs some regulatory compliance. it automated document analysis, ensuring compliance with legal, requirement and deadline, while, maintaining the accuracy. It also does improves, their legal research. OCR again improves legal research by making historical case documents and legal reference easier to find and access, which again improves a lot of time for, The fourth sector, which is our education technology, right? And, what it does there is the very nice feature that, OCI provides, to them is converting, speech, text to speech, right? And it, what it does for, the visual impairments. Impaired people. it's a boom to them, right? it gives them high, accurate and, whatever they, because they can't go through the documents, provides the speech, whatever they want to go through. Another use case, also is listening to books while driving. So OCF plays very important role in, in those sort of technologies. and the second, key factor is digital research, repository. OCR helps convert archive documents into searchable digital formats. it may making the educational resources easily available to the students and researchers. the third, important point here is the, it also auto automate the assessment. So OCR, that's it, streamlined the grading of the handwritten assignments, reducing administrative work for educators, and speeding up feedback for students. they can easily, the, the time that is, needed for an as assessment to be done is greatly reduced. which is very beneficial for all the, students and the researchers, coming to the cross sector analysis, right? We have gone through all the four sectors, healthcare, financial services. We've gone through the legal and the education. I. As you see in this particular table, right? all the sectors are broken down into what is the percentage of profic, processing efficiency, the document accuracy, what's the, percentage and time reduction, digital conversion rate, and the user satisfaction. So let's, let's go through what this table, states, right? And then we'll follow by what insights we can derive from it, right? now if you see the first one, which is a healthcare one, healthcare sector, it has a very strong, document accuracy of 98.2% and solid, pro, what you call as, processing efficiency, which is only 85% with a time reduction of 72.5%. coming to our second sector, which was financial. It leads in performing with the highest document accuracy, which is a nine, which is 99.1%, and pro processing efficiency of 92 7, along with the, time reduction off, 76.8%. coming to the legal one, it has a good performance where's 97.8% document accuracy with, and, 88.4% processing efficiency. It has some, little lower, time reduction, which is 70.2%, but very much acceptable. coming to education, it has lower processing efficiency of 82.6%, but very high user satisfaction. If you see there, it's an 89%, and, 65%, time reduction. So what we can derive from this particular table right across various sectors. financial services excel in most categories. That's the second rule that we can see. we also see that healthcare, shows strong document accuracy, which is used to diagnose, diagnose a prescription and whatnot. coming to education, it has very high satis, satisfaction despite low efficiency. which works in favors of edu, favor of education sector and all the sector experience, significant time saving compared to manual process. going to the next page, which is, what is the future development and integration challenges that OCR, has, right? So it's been shown, like the deep learning integration. it's currently deep learning machine learning. It's, it's in demand. it's going on. There's a strong acceptance rate and it's also with, in combination with OCR. It is it's gonna improve the feature extraction and pattern recognition, which will again enhance. everything in OCR, in all the all sectors. and regarding the second future development that we see is a multilingual recognition, handling diverse character sets and writing system. Third one is a edge computing deployment, which is optimizing models for devices and limited resources, regarding, data quality management. It has better, it'll be better people processing for handling that document variation. the fifth one is a security framework protected how important it is in all sector. We are, we, everyone is aware of, so it pro provide the protecting, sensitive information, right? So what we derive from it, right? What are the key insights, regarding the future? as we see here, right? OCR technology advancing, quickly, driven by AI and, deep learning challenges include, know, balancing, security, accuracy and scalability. which, along with ai, it's gonna improve a lot. Future, OCS system will go beyond character recognition to include contextual understanding. Multi-format processing and real time analytics. how important is analytics in all the industries? I don't think I have to explain, to anyone, but, so it's gonna be a really great combination of OCR with, there are deep learning, multilingual and machine, machine learning in the future. And these advances will revolutionize, in our digital sector, operations in sectors like healthcare, again, finance, legal and education, which again, which reducing the manual work and unlocking new insights. here I come to an end of the, this, my presentation. thank you for listening in. hope you gain some insight about OCR, and happy learning. Thank you. Thank you everyone.