Computer Vision Based Enterprise Process Automation in Multi- Cloud Environments with Data Protection Governance

Abstract

The increasing adoption of multi-cloud environments has created opportunities for enterprises to automate documentintensive,
inspection-oriented, and operational processes using artificial intelligence. Computer vision provides capabilities
for extracting information from images, scanned documents, video streams, and visual interfaces, enabling organizations to
automate tasks that traditionally require manual interpretation. However, deploying computer vision across multiple cloud
platforms introduces significant challenges involving data privacy, regulatory compliance, interoperability, model governance,
identity management, and secure data movement. This paper proposes a computer vision-based enterprise process automation
framework that integrates multi-cloud orchestration with data protection governance. The proposed framework combines
image preprocessing, optical character recognition, object detection, document classification, visual anomaly detection,
workflow automation, and policy-based data governance. Sensitive visual information is protected through data classification,
encryption, access control, minimization, retention management, and controlled cross-cloud data transfer. The research
methodology adopts a design-science and experimental approach involving requirements analysis, multi-cloud architecture
design, computer vision model development, governance-policy integration, workflow automation, and empirical evaluation.
The framework is evaluated using accuracy, processing time, automation rate, privacy-control effectiveness, data-transfer
overhead, compliance traceability, and workflow reliability. The study aims to demonstrate how computer vision can improve
enterprise process efficiency while maintaining appropriate governance over sensitive visual data. The proposed approach
provides a foundation for trustworthy visual automation across distributed cloud infrastructures while balancing operational
scalability, model performance, security, privacy, and regulatory requirements.