AI Vision
Oracle Cloud Infrastructure (OCI) Vision is an AI service for performing deep-learning–based image analysis at scale. With prebuilt models available out of the box, developers can easily build image recognition and text recognition into their applications without machine learning (ML) expertise. For industry-specific use cases, developers can automatically train custom vision models with their own data. These models can be used to detect visual anomalies in manufacturing, extract text from documents to automate business workflows, and tag items in images to count products or shipments.
Oracle Cloud Infrastructure Vision features
Detect objects and classify images
OCI Vision can classify images into thousands of categories to simplify common digital asset management scenarios or identify items that need attention, such as overgrowth near a power line. Developers can also identify and localize objects in images to automate counting of common items, such as packages and vehicles.
Detect text from scanned documents, mobile uploads, PDFs, and more
OCI Vision automatically extracts word- and line-level text from documents that are rotated, tilted, and shaded to support real-world document quality issues often found in expense processing and customer onboarding scenarios.
Extract structure from documents
Automatically identify and extract table structure from documents including the row and column location of each cell within your table. For specific documents, such as receipts, OCI Vision can identify key-value pairs, such as merchant name, total, tax, and tip.
Categorize documents
Identify and classify documents into common categories, such as invoice, receipt, and resume. Common applications include expense processing and enhanced search and retrieval of documents.
Train and manage models effortlessly
In addition to gaining access to pretrained models, developers can create custom models without managing custom model infrastructure. In addition to APIs, the OCI Vision user interface helps developers upload data and train an image classification or object detection model, review custom model metrics, and organize models into projects. This makes creating highly accurate, custom-trained computer vision models accessible to everyone.
Computer vision use cases
-
Gain insights from visual data
Automatically extract textual or visual information from images and use that to power analytic applications. Combine visual information from images with additional inputs, such as product sales history and customer reviews to gain a holistic sense of trends, such as product availability and popularity.
-
Automate back-office tasks
Classify documents, detect tables, and extract required information from documents, such as receipts, to automate business workflows, including employee expense reporting, spending compliance, customer loyalty programs, and reimbursement.
-
Digital asset management
Enrich image-based files with metadata, including document type, text, and objects for better indexing and retrieval in a digital asset management system or larger data warehouse. Easily retrieve images for uses ranging from intelligent search to retail management.
-
Detect visual anomalies
Classify products or equipment as standard or defective based on visual appearance, such as discoloration, tear, rust, deformity, or breaks. Automate the detection of defective materials to flag the need for repairs.
-
Monitor scenes
Detect whether vegetation is growing in the surveillance image of a power line or if trucks are available at a lot for delivery by locating objects and entire scenes in images. Localize where particular objects are in a scene or classify the entire image.
OCI Vision pricing
Image Analysis
Product |
Unit price |
Unit |
Oracle Cloud Infrastructure - Vision - Image Analysis - 1,000 Transactions (First 5,000 Transactions) |
1000 Transactions | |
Oracle Cloud Infrastructure - Vision - Image Analysis - 1,000 Transactions (Greater than 5,000 Transactions) |
1000 Transactions |
OCR
Product |
Unit price |
Unit |
Oracle Cloud Infrastructure - Vision - OCR - 1,000 Transactions (First 5,000 Transactions) |
1000 Transactions | |
Oracle Cloud Infrastructure - Vision - OCR - 1,000 Transactions (Greater than 5,000 Transactions) |
1000 Transactions |
Document Properties
Product |
Unit price |
Unit |
Oracle Cloud Infrastructure - Vision - Document Properties - 1,000 Transactions (First 5,000 Transactions) |
1000 Transactions | |
Oracle Cloud Infrastructure - Vision - Document Properties - 1,000 Transactions (Greater than 5,000 Transactions) |
1000 Transactions |
Document Extraction
Product |
Unit price |
Unit |
Oracle Cloud Infrastructure - Vision - Document Extraction - 1,000 Transactions (First 5,000 Transactions) |
1000 Transactions | |
Oracle Cloud Infrastructure - Vision - Document Extraction - 1,000 Transactions (Greater than 5,000 Transactions) |
1000 Transactions |
Custom Training
Product |
Unit price |
Unit |
Oracle Cloud Infrastructure - Vision- Custom Training - Training Hour (First 15 Training Hour) |
Training hour | |
Oracle Cloud Infrastructure - Vision- Custom Training - Training Hour (Greater than 15 Training Hour) |
Training hour | |
| Machine used for training may vary based on availability. | ||
OCI Vision monthly pricing examples
General digital asset management
Classify one million images, with an average of 100,000 new images added each month.
Initial tagging: 1 million images x ($0.25/1000) images = $250
Every month afterward: 100,000 images x ($0.25/1000) images= $25/month
Visual anomaly detection
Train custom object detection model for two hours, and review 1,000 images a day (approximately 30,000 images per month) to look for potential defects.
Initial training: 2 hours x ($1.5/hour) = $3
Monthly custom image analysis: 30,000 images x ($0.25/1000) images=$7.5
Downstream NLP
Extract terms and timing schedule for commitments from 1000 contracts per month, with each contract at approximately 50 pages.
Monthly cost for OCR SKU: 1000 contracts x (50 pages/contract) x ($1/1000) images = $50
Receipt processing
Extract text and key-value pairs from approximately 50,000 receipts per month to complete returns, with each receipt at 1 page long.
Monthly cost for document extraction SKU: 50,000 documents x 1 page per doc x ($10/1000 pages) = $500/month
