Search & Discovery

CoreViz provides powerful search capabilities that go far beyond simple filename matching. This guide covers all the ways you can find and discover content in your visual datasets.

Overview

CoreViz offers multiple search modalities that work together to help you find exactly what you're looking for:

• Natural Language Search: Describe what you're looking for in plain English
• Ask AI: Have a conversation with the CoreViz AI agent to find exactly what you need
• Visual Similarity Search: Find images similar to a reference image
• Object-Based Search: Search for specific objects, people, or scenes
• Metadata Filtering: Filter by file properties, tags, and custom attributes
• Advanced Filters: Combine multiple criteria for precise results

Natural Language Search

The most intuitive way to search your content is using natural language descriptions.

Basic Text Search

Simply type what you're looking for in the search box:

"red car in parking lot"
"person wearing blue shirt"
"product with visible logo"
"sunset over mountains"
"dog playing in water"

Search Tips

Be Descriptive:

• ✅ "red sports car in empty parking lot"
• ❌ "car"

Include Context:

• ✅ "person sitting at desk with computer"
• ❌ "person"

Use Specific Details:

• ✅ "woman wearing red dress holding umbrella"
• ❌ "woman with umbrella"

Chat Agent Search

The most powerful way to search your content is through conversation with CoreViz's AI agentic chat. Simply talk to the AI as you would to a human assistant, and it will help you find exactly what you're looking for.

Example conversation with CoreViz's AI chat agent for finding product photos

How Chat Search Works

Instead of typing search queries, you have a natural conversation with the AI:

You: "I need to find all the product photos we took last month that show our new red sneakers"

AI: "I'll help you find those photos. Let me search for images from last month that contain red sneakers. I found 47 results. Would you like me to filter by specific angles or show you the highest quality ones first?"

You: "Show me the ones with the best lighting and make sure they're from the main product shoot"

AI: "Perfect! I've filtered to 12 high-quality images from your main product shoot. These all have excellent lighting and show the red sneakers clearly. Would you like me to organize them by angle or color variation?"

Conversation Examples

Product Discovery:

You: "Find me all the images where customers are using our product"
AI: "I found 23 images showing customer usage. Would you like me to categorize them by product type or usage scenario?"

You: "Show me the ones with the kitchen products"
AI: "Here are 8 images of customers using kitchen products. I can see they're using the blender, food processor, and coffee maker."

Content Analysis:

You: "I need to find all the images that might have copyright issues"
AI: "I'll scan your content for potential copyright concerns. I found 15 images that contain text, logos, or artwork that might need review. Would you like me to show them organized by risk level?"

You: "Show me the high-risk ones first"
AI: "Here are 5 high-risk images that contain visible brand logos or copyrighted artwork. I recommend reviewing these before use."

Research & Investigation:

You: "Find all the surveillance footage from the main entrance between 2-4 PM yesterday"
AI: "I found 12 video clips from the main entrance during that time period. Would you like me to highlight any unusual activity or focus on specific individuals?"

You: "Show me if there were any people in red clothing"
AI: "I found 3 people wearing red clothing during that time period. Here are the relevant clips with timestamps."

Chat Search Benefits

Natural Interaction:

• Ask questions in plain English
• Refine searches through conversation
• Get contextual suggestions and insights

Intelligent Assistance:

• AI understands context and intent
• Suggests related searches and filters
• Explains why certain results were found

Iterative Refinement:

• Start broad and narrow down through conversation
• Get suggestions for better search terms
• Discover content you didn't know existed

Best Practices for Chat Search

Be Conversational:

• ✅ "I'm looking for photos from our summer campaign that show people having fun"
• ❌ "summer campaign photos people fun"

Provide Context:

• ✅ "Find me the product shots we took for the holiday catalog last year"
• ❌ "holiday product shots"

Ask Follow-up Questions:

• ✅ "Can you show me the ones with the best composition?"
• ✅ "Are there any similar images I might have missed?"
• ✅ "Can you organize these by date or quality?"

Use Natural Language:

• ✅ "I need to find all the images that might be too dark for our website"
• ✅ "Show me everything from the trade show that has our logo visible"
• ✅ "Find me the best examples of our product being used in real life"

Advanced Chat Features

Multi-Modal Queries:

You: "Find images similar to this one but with better lighting"
[Upload reference image]
AI: "I found 8 similar images with improved lighting. Would you like me to show you the ones that also match your brand colors?"

Complex Workflows:

You: "I need to create a presentation about our product evolution. Can you find images from each major version and organize them chronologically?"
AI: "I'll help you build that presentation. I found images from 5 major product versions. Let me organize them by release date and highlight the key visual changes between versions."

Analytical Insights:

You: "What patterns do you see in our most successful product photos?"
AI: "Based on your high-performing images, I notice they typically feature: 1) Natural lighting, 2) Clean backgrounds, 3) Multiple angles, and 4) Lifestyle contexts. Would you like me to find more images that match these patterns?"

Visual Similarity Search

Find visually similar content by uploading a reference image or selecting from your existing content.

CoreViz search interface showing visual similarity search in action

How It Works

1. Upload Reference: Upload an image or select from your dataset
2. AI Analysis: CoreViz analyzes visual features (colors, shapes, objects, composition)
3. Similarity Matching: Find images with similar visual characteristics
4. Ranked Results: Results ordered by visual similarity score

Use Cases

Product Matching:

• Find similar products in your catalog
• Identify duplicate or near-duplicate images
• Discover related items for recommendations

Content Discovery:

• Find images with similar composition or style
• Locate content with matching color schemes
• Discover related visual themes

Quality Control:

• Identify variations of the same product
• Find images that don't match your brand style
• Detect potential copyright issues

Similarity Search Tips

Good Reference Images:

• Clear, high-quality images
• Well-lit and in focus
• Representative of what you're looking for
• Minimal background distractions

Adjusting Results:

• Use similarity threshold slider
• Combine with text search for refinement
• Filter by metadata for additional context

Object-Based Search

Search for specific objects, people, or scenes detected by AI.

Detected Objects

CoreViz automatically detects and indexes:

People & Animals:

• Person, face, hand, eye
• Dog, cat, bird, horse
• Baby, child, adult

Vehicles:

• Car, truck, bus, motorcycle
• Bicycle, airplane, boat
• Train, helicopter

Objects:

• Phone, computer, book
• Chair, table, bed
• Food, drink, bottle

Scenes:

• Indoor, outdoor, street
• Beach, mountain, forest
• Office, kitchen, bedroom

Object Search Examples

"find all images with cars"
"videos containing people"
"images with text or signs"
"photos of food or drinks"
"pictures with animals"

Metadata Filtering

Filter your search results using file properties and custom metadata.

File Properties

Basic Properties:

• File type (image/video)
• File format (JPEG, PNG, MP4, etc.)
• File size range
• Upload date range
• Processing status

Image Properties:

• Dimensions (width × height)
• Color space
• Compression quality
• Orientation

Video Properties:

• Duration
• Frame rate
• Resolution
• Bitrate

Search Performance

Optimization Tips

Efficient Queries:

• Be specific rather than broad
• Use filters to narrow results
• Combine text and visual search
• Save common searches

Large Datasets:

• Use date filters for recent content
• Filter by dataset or entity
• Use object-based search for specific items
• Consider batch processing for bulk operations

Search Limits

Result Limits:

• Maximum 1000 results per search
• Use pagination for large result sets
• Apply additional filters to narrow results

Query Complexity:

• Complex queries may take longer
• Use simpler queries for better performance
• Break complex searches into multiple steps

Search Use Cases

E-commerce

Product Discovery:

"red dress size medium"
"shoes with laces"
"electronics under $100"

Inventory Management:

"products without tags"
"images needing approval"
"duplicate products"

Security & Surveillance

Incident Investigation:

"person in red shirt yesterday"
"vehicle license plate ABC123"
"activity between 2-4 PM"

Pattern Analysis:

"repeated visitors"
"unusual activity"
"crowd formation"

Content Management

Asset Organization:

"marketing images 2024"
"high-resolution photos"
"branded content"

Quality Control:

"blurry images"
"low-quality videos"
"inappropriate content"

Troubleshooting Search

Common Issues

No Results Found:

• Check spelling and grammar
• Try broader search terms
• Verify content is processed and indexed
• Check date and filter settings

Too Many Results:

• Add more specific terms
• Use additional filters
• Narrow date ranges
• Use object-based search

Slow Search Performance:

• Simplify complex queries
• Use filters to reduce scope
• Check network connection
• Contact support for large datasets

Search Tips

Best Practices:

• Start broad, then narrow down
• Use multiple search types together
• Save successful searches
• Review search history for patterns

Advanced Techniques:

• Use chat agent for complex, conversational searches
• Combine text and visual search
• Use metadata for precise filtering
• Leverage object detection results
• Create search workflows

Next Steps

Now that you understand search and discovery:

• Explore Workflows: Discover AI workflows for specialized analysis
• Bulk Operations: Learn bulk operations for large-scale tasks
• Integrations: Connect with integrations and external tools
• Performance: Optimize with performance tips