- We offer certified developers to hire.
- We’ve performed 1500+ Web/App/eCommerce projects.
- Our clientele is 1000+.
- Free quotation on your project.
- We sign NDA for the security of your projects.
- Three months warranty on code developed by us.
Organizations across industries continue to deal with massive volumes of documents in both physical and digital formats. These documents include invoices, purchase orders, contracts, bank statements, receipts, shipping documents, legal paperwork, identification records, and medical reports. Traditionally, extracting data from such documents required manual processing, where employees would read documents and enter the relevant information into digital systems.
Manual document processing is time consuming, prone to human error, and costly when organizations must handle thousands or millions of documents every month. Optical Character Recognition technology has emerged as a powerful solution that helps organizations automate document processing and convert unstructured document images into structured digital data.
OCR software development services focus on building intelligent systems capable of extracting text and data from scanned documents, images, and PDFs automatically. OCR technology analyzes visual characters within document images and converts them into machine readable text that can be processed by digital applications.
Modern OCR systems go far beyond basic character recognition. With the integration of artificial intelligence, machine learning, and computer vision technologies, OCR platforms can analyze document structures, identify key data fields, classify document types, and automate document processing workflows.
For example, when a company receives supplier invoices in PDF or scanned format, an OCR system can automatically extract invoice numbers, supplier names, dates, and payment amounts. This information can then be transferred directly into accounting systems or enterprise resource planning platforms without manual intervention.
OCR software development services are widely used in industries such as banking, insurance, healthcare, logistics, retail, legal services, and government operations. Financial institutions use OCR systems to process loan applications and financial forms. Healthcare organizations use OCR solutions to digitize patient records and insurance claims.
Logistics companies use OCR technology to process shipping documents and customs forms. Legal organizations use OCR platforms to analyze contracts and legal records quickly.
Developing advanced OCR systems requires expertise in artificial intelligence, image processing, natural language processing, and enterprise software development. Technology companies specializing in OCR and AI solutions help organizations implement document automation platforms tailored to their business needs.
Organizations such as <a href=”https://www.abbacustechnologies.com/”>Abbacus Technologies</a> provide OCR software development services that enable businesses to automate document workflows and digitize paper based records. These solutions combine machine learning models, scalable cloud infrastructure, and enterprise integrations to create intelligent document automation platforms.
Understanding how OCR software development works helps organizations adopt automation technologies that improve efficiency and reduce operational costs.
OCR document automation systems analyze scanned documents and convert the visual text contained within those documents into machine readable digital text. These systems combine image processing algorithms, text recognition models, and natural language processing technologies to interpret document content.
The process begins when a document image is uploaded to the system. Documents may be captured through scanners, mobile device cameras, or digital file uploads such as PDF documents.
Once the document image is uploaded, it is transmitted to the OCR processing engine. The first stage of processing involves image preprocessing.
Document images often contain distortions caused by scanning or photography. These distortions may include skewed alignment, background noise, shadows, or uneven lighting. Image preprocessing algorithms correct these issues by adjusting brightness, removing noise, correcting skew angles, and enhancing contrast.
After preprocessing, the system performs text detection. Computer vision algorithms analyze the document image and identify regions that contain textual information. These algorithms separate text from graphics, logos, tables, and background elements.
Once text regions are detected, OCR algorithms analyze the characters within those regions. The OCR engine recognizes letters, numbers, and symbols and converts them into machine readable text.
Modern OCR engines use deep learning models to improve recognition accuracy even when documents contain complex fonts, low resolution text, or handwritten annotations.
After extracting the text, natural language processing algorithms analyze the textual content to identify meaningful information. For example, the system may identify names, addresses, dates, invoice numbers, and monetary values.
Document classification models may also categorize documents into specific types such as invoices, receipts, contracts, bank statements, or identification documents.
Once the relevant information is extracted, the system converts the data into structured formats such as database records or JSON objects. This structured data can then be integrated into enterprise systems such as accounting platforms, CRM software, or document management systems.
OCR document automation systems therefore transform unstructured document images into structured digital data that can be used in automated workflows.
OCR software development services rely on several advanced technologies that work together to automate document processing.
Artificial intelligence and machine learning algorithms form the foundation of modern OCR systems. Machine learning models are trained on large datasets of document images to recognize characters and document patterns.
Computer vision algorithms analyze document images to detect text regions and understand document layouts.
Optical character recognition engines convert visual text into machine readable text.
Natural language processing models analyze extracted text and identify key entities such as names, addresses, and financial values.
Document classification models categorize documents into different types based on visual and textual features.
Table extraction algorithms identify tabular data within documents and convert it into structured spreadsheet formats.
Cloud computing infrastructure supports large scale document processing and AI model training.
Enterprise integration frameworks allow OCR systems to connect with business applications such as ERP platforms and document management systems.
Data analytics platforms analyze extracted document data to generate insights about business operations.
The integration of these technologies enables businesses to build intelligent OCR platforms that automate document processing tasks.
Modern OCR software development services provide several features designed to streamline document automation workflows.
Automated text extraction allows organizations to convert scanned documents into searchable digital text.
Document classification systems identify document types and route them to appropriate processing workflows.
Data extraction tools identify key information fields such as names, addresses, invoice numbers, and financial values.
Table recognition features extract tabular data from documents and convert it into structured spreadsheet formats.
Handwritten text recognition enables OCR systems to interpret handwritten forms and notes.
Multi language support allows OCR platforms to process documents written in different languages.
Integration capabilities allow extracted data to be transferred into enterprise systems automatically.
Analytics dashboards provide insights into document processing performance and operational efficiency.
OCR based document automation platforms provide numerous advantages for organizations handling large volumes of documents.
Improved efficiency allows businesses to process documents quickly without relying on manual data entry.
Faster document processing enables organizations to handle large document volumes in real time.
Reduced human error improves data accuracy and reliability.
Cost savings result from reducing labor intensive document processing tasks.
Enhanced search capabilities allow organizations to locate document information quickly.
Scalable systems enable businesses to expand operations without increasing administrative workloads.
Improved compliance and audit readiness ensure that documents are stored and managed properly.
OCR document automation technologies support a wide range of applications across industries.
Financial institutions use OCR platforms to process loan applications and bank statements.
Insurance companies use OCR systems to analyze claim documents and policy forms.
Healthcare organizations use OCR solutions to digitize patient records and medical forms.
Retail companies use OCR technology to process receipts and invoices.
Legal organizations use OCR platforms to analyze contracts and legal documents.
Government agencies use OCR systems to digitize public records and administrative forms.
Logistics companies use OCR software to process shipping documents and customs paperwork.
These applications demonstrate how OCR software development services are transforming document management and enterprise automation.OCR software development services for document automation represent a major advancement in enterprise digital transformation. By combining optical character recognition technology with artificial intelligence and computer vision, organizations can automate document processing workflows and extract valuable information from document images.
OCR based document automation platforms enable businesses to digitize documents, reduce manual workloads, and improve operational efficiency.
As artificial intelligence technologies continue to evolve, OCR software solutions will become even more powerful, enabling organizations to build intelligent document management systems and automate complex administrative processes.
Developing OCR software for document automation requires a robust and scalable architecture capable of processing large volumes of document images while maintaining high accuracy and reliability. Organizations often deal with documents coming from multiple sources including scanners, mobile devices, emails, and enterprise systems. A well designed architecture ensures that these documents can be captured, analyzed, and converted into structured data efficiently.
The architecture of an OCR document automation platform typically begins with the document acquisition layer. This layer is responsible for capturing documents from various input sources. Documents may be uploaded as scanned images, photographed using mobile devices, or submitted as digital PDF files. Some organizations also integrate OCR systems with email servers or enterprise applications that automatically send documents to the processing platform.
Once the document is captured, it enters the data ingestion layer. This component handles the secure transfer of document files into the OCR processing system. Application programming interfaces allow different business systems, mobile apps, and web platforms to submit documents directly to the OCR engine.
After ingestion, the document image moves to the preprocessing stage. Document images captured from scanners or cameras often contain imperfections that affect text recognition accuracy. These imperfections may include skewed alignment, shadows, low contrast, noise, or uneven lighting.
Image preprocessing algorithms improve the quality of the document image by correcting skew angles, removing noise, adjusting brightness and contrast, and standardizing image resolution. These corrections ensure that the document is properly aligned for further analysis.
Once the document image is enhanced, the system performs layout detection. Computer vision algorithms analyze the structure of the document and identify different regions such as text blocks, tables, headers, footers, form fields, and logos. Layout detection is important because it helps the system understand the organization of the document.
After identifying layout structures, the document enters the text recognition stage. OCR engines analyze the detected text regions and convert visual characters into machine readable text. Modern OCR systems use deep learning models to recognize characters even when documents contain complex fonts or degraded text.
Once the text is extracted, the system proceeds to text interpretation. Natural language processing algorithms analyze the extracted text and identify key information such as names, addresses, invoice numbers, dates, product descriptions, and financial values.
Entity recognition models categorize extracted text elements into meaningful data fields. For example, an invoice document may contain fields such as supplier name, invoice date, invoice number, and total amount.
Document classification models also play an important role in OCR systems. These models analyze the document image and categorize it into predefined types such as invoices, receipts, contracts, bank statements, or identification documents.
After extracting relevant information, the OCR system converts the results into structured formats such as JSON, XML, or database entries. These structured data formats allow the extracted information to be integrated directly into enterprise software systems.
The application layer provides user interfaces that allow employees and administrators to interact with the OCR system. Users can upload documents, review extracted data, validate results, and export information to other business applications.
Cloud computing infrastructure supports the entire OCR processing pipeline. Cloud platforms provide scalable computing resources that allow organizations to process large volumes of documents simultaneously without performance degradation.
Data storage systems maintain document images, extracted text, and processing history. These records can be used for auditing, compliance, and machine learning improvements.
Security layers protect sensitive document information through encryption protocols, authentication mechanisms, and role based access control.
This architecture enables OCR software platforms to automate document processing workflows and support enterprise scale document management systems.
Deep learning models are essential components of modern OCR software. These models enable systems to recognize characters, interpret document layouts, and extract meaningful information from document images.
Convolutional neural networks are widely used in OCR systems because they are effective at detecting visual patterns in images. These networks analyze document images and identify text blocks, characters, and layout structures.
Text detection models identify regions within document images that contain textual information. These models help separate text from graphics and background elements.
Character recognition models convert detected characters into machine readable text. Deep learning based OCR models can recognize text across different fonts, languages, and image conditions.
Natural language processing models analyze extracted text and identify entities such as names, dates, addresses, and financial values.
Document classification models analyze visual and textual features to categorize documents into specific types.
Table extraction models identify tabular data within documents and convert it into structured spreadsheet formats.
Continuous model training ensures that OCR systems improve accuracy as they process new document formats and layouts.
OCR document automation systems must integrate seamlessly with enterprise software platforms to deliver maximum operational benefits.
Enterprise resource planning systems manage financial transactions, procurement operations, and inventory management. OCR platforms can extract invoice data and send it directly to ERP systems for automated accounting processes.
Customer relationship management systems store customer information and communication records. OCR systems can extract customer data from forms and integrate it with CRM platforms.
Human resource management systems maintain employee records and payroll information. OCR solutions can process employee documents such as resumes, identification forms, and employment contracts.
Document management systems store digital documents and enable document retrieval. OCR platforms enhance these systems by converting scanned documents into searchable text.
Technology companies specializing in AI and OCR development, including Abbacus Technologies, build document automation solutions that integrate seamlessly with enterprise software environments.
High quality datasets are essential for training OCR models. These datasets contain large collections of document images representing different document types and formats.
Before these datasets can be used for machine learning training, they must undergo annotation. Annotation involves labeling document images with information about text regions, layout structures, and data fields.
Annotators mark areas containing text blocks, tables, form fields, and signature sections. These annotations help machine learning models learn how to interpret document layouts.
Domain experts may assist in labeling specific data fields such as invoice numbers, tax values, or contract clauses.
Accurate annotations ensure that OCR models learn meaningful patterns from the training data.
Data augmentation techniques are often used to expand document datasets. Images may be rotated, blurred, or distorted to simulate real world scanning conditions.
Dataset management systems store document datasets and organize them efficiently for training and evaluation.
OCR document automation platforms must implement strong security and data management practices to protect sensitive information.
Documents processed by OCR systems often contain confidential data such as financial records, personal identification details, legal agreements, or medical reports.
Encryption protocols protect document images during transmission between users and processing servers.
Access control mechanisms ensure that only authorized personnel can view or modify document data.
Data analytics platforms analyze document processing activities to generate insights about workflow efficiency and document processing performance.
Responsible data management practices ensure that OCR document automation systems operate securely while supporting enterprise scale operations.
Developing OCR software for document automation requires a structured development approach that combines expertise in computer vision, artificial intelligence, natural language processing, and enterprise software engineering. Organizations rely on OCR systems to process thousands or even millions of documents efficiently, which means that these systems must be accurate, scalable, and capable of handling diverse document formats. The development process therefore involves several stages, including requirement analysis, dataset preparation, model training, system integration, and continuous optimization.
The development process begins with requirement analysis and workflow evaluation. During this stage, developers collaborate with business stakeholders, IT teams, and operational departments to understand how documents are currently handled within the organization. Companies across industries process various document types such as invoices, receipts, bank statements, contracts, shipping documents, identification forms, and application forms. Each of these document types has unique layouts and information fields that must be extracted.
The goal of this phase is to identify the document types the OCR system must support and determine what information needs to be extracted from each document. For example, an invoice processing solution may require the extraction of supplier names, invoice numbers, invoice dates, tax values, line items, and total amounts. A banking document processing system may need to extract account numbers, transaction details, and statement dates.
Understanding these requirements helps developers design the OCR system architecture and determine which AI models will be needed for document recognition and data extraction.
Once the requirements are clearly defined, the next stage involves dataset collection. OCR systems require large datasets of document images for training machine learning models. These datasets must represent the types of documents that the system will process in real world environments.
Document datasets may include scanned invoices, photographed receipts, bank statements, contracts, application forms, and identification documents. It is important that the dataset includes documents captured under different conditions such as varying lighting environments, scanning resolutions, and camera angles. This diversity ensures that the OCR system can perform reliably in practical situations.
After collecting the document dataset, the images must undergo annotation. Annotation is the process of labeling document images with information about text regions, layout structures, and specific data fields. Annotators mark areas within documents that contain textual information, tables, form fields, and graphical elements.
Domain experts may assist in labeling key data fields within documents. For example, in invoice processing datasets, experts may identify fields such as invoice numbers, payment terms, and product descriptions. These annotations provide the ground truth data used for training machine learning models.
Once the annotated dataset is prepared, developers proceed to the AI model development stage. Machine learning engineers design deep learning architectures capable of recognizing text and document structures.
Computer vision models analyze document images and detect regions containing text, tables, or other structural components. These models help the system understand the spatial organization of document content.
OCR engines are then trained to recognize characters within the detected text regions. Modern OCR systems rely on deep learning models that can recognize characters across various fonts, languages, and document conditions.
Natural language processing models analyze the extracted text and identify meaningful entities such as names, addresses, dates, product descriptions, and financial values. These models help convert raw text into structured information that can be used by business applications.
Machine learning models also perform document classification tasks. The system learns to categorize documents into predefined types such as invoices, receipts, contracts, or bank statements. This classification step ensures that documents are routed to the appropriate processing workflows.
During training, annotated document images are fed into neural network models. The system generates predictions about the document content and compares those predictions with the annotated ground truth data. If errors occur, the model adjusts its internal parameters through iterative training cycles.
Training OCR models requires significant computational resources because document datasets can contain millions of images. Cloud based machine learning platforms and graphics processing units are typically used to accelerate the training process.
After training is complete, the OCR system undergoes validation and testing. Validation datasets contain document images that were not used during training and are used to evaluate the model’s ability to recognize new documents accurately.
Testing also involves evaluating the system using real world documents captured from enterprise workflows. This ensures that the system performs reliably across different document formats and scanning conditions.
Once the OCR models demonstrate high accuracy, developers integrate the system with enterprise applications. APIs allow business systems such as accounting software, CRM platforms, and document management systems to send documents to the OCR engine and receive extracted data automatically.
For example, when a supplier invoice is uploaded to the system, the OCR platform analyzes the document, extracts the relevant information, and sends the structured data to the accounting system.
Before full deployment, organizations often conduct pilot programs with selected departments. These pilot implementations allow teams to evaluate system performance and identify areas for improvement.
Technology companies specializing in OCR and AI development, including Abbacus Technologies, follow structured development methodologies to build enterprise grade OCR platforms that integrate seamlessly with document automation workflows.
Although OCR technology has advanced significantly, developing reliable OCR systems still presents several challenges.
One major challenge is the diversity of document layouts. Different organizations use unique document formats, fonts, and layouts, which makes it difficult for OCR systems to interpret them consistently.
Another challenge involves poor document image quality. Scanned documents may contain noise, blurred text, low resolution characters, or distortions that reduce recognition accuracy.
Handwritten text recognition is also difficult because handwriting styles vary widely among individuals.
Language diversity can also create challenges for OCR systems operating in global environments where documents may appear in multiple languages.
Despite these challenges, advancements in deep learning and computer vision technologies continue to improve OCR accuracy and reliability.
Organizations implementing OCR solutions often choose between generic OCR tools and custom built OCR platforms tailored to their business needs.
Generic OCR tools can convert document images into text but often lack the ability to interpret document structures or extract specific data fields.
Custom OCR platforms are designed to understand document layouts and extract structured data relevant to specific business workflows.
Custom solutions can be trained using organization specific document datasets, improving accuracy for specialized document formats.
Integration capabilities are another advantage of custom development. OCR systems can integrate directly with enterprise software systems such as ERP platforms, document management systems, and workflow automation tools.
Although generic OCR tools may provide basic functionality, custom OCR platforms offer greater flexibility and automation capabilities for enterprise document processing.
Developing OCR software for document automation involves several cost factors that organizations must consider.
Dataset preparation is one of the most significant costs because annotating document images requires skilled data labeling teams and domain expertise.
Computational infrastructure is another major cost factor. Training deep learning models on large document datasets requires powerful GPU hardware or cloud based machine learning platforms.
Software development costs include building OCR algorithms, document processing interfaces, integration APIs, and analytics dashboards.
Cloud infrastructure costs may arise from storing document images and processing large volumes of document analysis requests.
Maintenance and model updates represent ongoing costs because OCR models must be retrained periodically to support new document formats.
Despite these costs, OCR document automation platforms provide substantial long term value by reducing manual workloads and improving operational efficiency.
OCR based document automation technologies are transforming business operations by eliminating manual document processing tasks.
Organizations can process documents faster and more accurately while reducing administrative workloads.
Business processes such as invoice processing, contract analysis, customer onboarding, and compliance documentation can be automated using OCR systems.
Structured document data enables organizations to perform advanced analytics and gain insights into operational activities.
By integrating OCR technology into enterprise workflows, businesses can streamline operations, improve productivity, and accelerate digital transformation initiatives.
Selecting the right technology partner is a critical step for organizations planning to implement OCR software for document automation. Since OCR platforms must analyze large volumes of documents, interpret complex layouts, and integrate with enterprise business systems, the development company must have strong expertise in artificial intelligence, computer vision, and enterprise software architecture.
One of the most important factors to evaluate when choosing an OCR development partner is their experience in building AI driven document processing solutions. OCR systems rely on machine learning models capable of recognizing characters, detecting document structures, and extracting meaningful information from document images. Developers must understand how to train neural networks using large document datasets and optimize them to perform accurately across different document formats and languages.
Another key factor is experience with enterprise system integration. OCR platforms often operate as part of a broader automation ecosystem that includes enterprise resource planning systems, accounting software, document management platforms, and customer relationship management tools. A development partner with integration expertise can ensure that extracted document data flows seamlessly into business applications and automated workflows.
Scalability is also an important consideration when selecting an OCR development company. Large organizations may process thousands or even millions of documents every month. The OCR platform must be capable of handling high document volumes while maintaining fast processing speeds and high recognition accuracy.
Security and compliance are also essential aspects of document automation systems. Documents often contain sensitive information such as financial records, personal identification data, legal agreements, or medical information. Development teams must implement strong encryption protocols, secure cloud infrastructure, and strict access control policies to protect confidential data.
User experience design also plays a major role in successful OCR platforms. Employees interacting with the system should be able to upload documents easily, review extracted information quickly, and validate data when necessary. Intuitive dashboards and workflow management interfaces improve user adoption and operational efficiency.
Long term support and maintenance services should also be considered when choosing a development partner. OCR models must be updated regularly as new document formats, languages, and layouts appear in business environments. Continuous improvement ensures that the OCR system remains accurate and reliable.
Organizations seeking specialized expertise in AI powered document automation often collaborate with experienced technology providers. Companies such as <a href=”https://www.abbacustechnologies.com/”>Abbacus Technologies</a> provide OCR software development services designed to help businesses automate document processing workflows and digitize paper based records. Their expertise in artificial intelligence, scalable cloud infrastructure, and enterprise integration enables organizations to deploy powerful document automation platforms.
Choosing the right development partner ensures that OCR solutions are built with the reliability, scalability, and security required for enterprise level document processing.
OCR software development offers numerous advantages for organizations seeking to automate document processing tasks.
One of the most significant benefits is improved operational efficiency. OCR systems eliminate the need for manual data entry by automatically extracting information from documents. This allows organizations to process large document volumes much faster than traditional manual workflows.
Faster document processing enables businesses to respond to operational tasks more quickly. For example, invoice processing times can be reduced significantly when OCR systems automatically extract financial information and transfer it to accounting systems.
Improved accuracy is another important advantage. Machine learning based OCR models can extract text consistently and reduce the errors that commonly occur in manual data entry processes.
Cost reduction is also a major benefit of OCR automation. By reducing manual document processing tasks, organizations can lower labor costs and improve resource allocation.
Enhanced search and retrieval capabilities are possible because OCR systems convert documents into searchable digital text. Employees can locate specific information within documents quickly using keyword searches.
Improved compliance and auditing capabilities are also achieved through structured document storage and digital record management.
The field of OCR and document automation is evolving rapidly as artificial intelligence technologies continue to advance. Several emerging trends are shaping the future of OCR software development.
One important trend is intelligent document understanding. These systems go beyond basic text extraction and analyze the semantic meaning of documents. AI models can understand relationships between different pieces of information and identify important context within documents.
Multilingual OCR technology is also improving. Modern OCR systems can process documents written in multiple languages and recognize characters across different alphabets.
Handwritten text recognition is becoming more accurate with the help of deep learning models. Advanced handwriting recognition systems can interpret handwritten notes and form entries with high reliability.
AI powered contract analysis platforms are also gaining popularity. These systems analyze legal documents automatically and identify clauses, risks, and compliance requirements.
Mobile OCR solutions are another emerging trend. Employees can capture document images using smartphones, and OCR systems can process these images instantly.
These innovations are expanding the capabilities of OCR technology and enabling organizations to automate increasingly complex document workflows.
OCR document automation systems must undergo continuous training and optimization to maintain high levels of performance and accuracy.
Business environments frequently introduce new document formats, templates, and layouts. OCR models must be retrained periodically to recognize these changes.
Continuous model training allows the OCR system to learn from new document images and improve recognition accuracy over time.
Validation processes ensure that OCR models perform consistently across different document types and scanning conditions.
Performance monitoring tools help organizations track metrics such as recognition accuracy, processing speed, and document processing volumes.
Software updates may introduce improved character recognition algorithms, enhanced document classification models, and better integration features.
Security updates are also critical for protecting sensitive document data and maintaining compliance with regulatory requirements.
Organizations that treat OCR systems as evolving platforms rather than static software can ensure long term reliability and continuous improvement.
OCR technology is being widely adopted across industries as organizations pursue digital transformation and automation strategies.
Financial institutions use OCR systems to automate loan application processing and document verification.
Insurance companies rely on OCR platforms to analyze claim documents and policy forms.
Healthcare organizations use OCR technology to digitize patient records and streamline administrative workflows.
Retail companies use OCR solutions to process receipts, invoices, and inventory documents.
Government agencies implement OCR platforms to digitize public records and improve administrative services.
Logistics companies use OCR technology to analyze shipping documents, bills of lading, and customs paperwork.
The increasing availability of cloud computing infrastructure and AI development frameworks has made OCR technology more accessible to organizations of all sizes.
As businesses continue to adopt automation technologies, OCR software development will play a crucial role in enabling intelligent document management systems.
OCR software development services for document automation represent a major advancement in enterprise digital transformation. By combining optical character recognition with artificial intelligence and computer vision technologies, organizations can automate the extraction of information from document images and streamline administrative processes.
OCR powered document automation platforms help businesses digitize documents, reduce manual workloads, and improve operational efficiency.
As artificial intelligence technologies continue to evolve, OCR systems will become even more powerful and capable, enabling organizations to build intelligent document processing solutions that support modern digital business operations.
Get cost, timeline & technical suggestions within 24 hours.
