pixolution Flow comes with pre-built, pre-trained modules.
Go beyond simple keyword searches. Use an image to search your collection and find the most relevant images with similar content.
DIGITAL ASSET MANAGEMENT VENDOR
Pixelboxx uses pixolution Flow as an integral part of its internal search engine and exposes the search via API. Using the Adobe Creative Plugin, designers can find image inspiration from the connected pixelboxx DAM directly from Photoshop and InDesign. This speeds up the creative workflow by avoiding context switches. The deep integration of image search into the pixelboxx DAM brings the search to where the user works.
Search results are ranked by their relevance. Based on the query parameters each document is scored representing the confidence and relevance.
Apply search filter to restrict searches to a specific sub-set of your collection. Multiple search criteria can be combined via logical operators (AND, OR, NOT).
Geo-spatial searches are supported to search within certain regions, locations or ranges. Results that are further away will be ranked lower.
Filter by ranges and dates. Search results can be faceted to group results in multitude of ways.
Combine different data types in one search query. Search for image and text at the same time.
Decouple analysis and indexing as required.
Just provide the image URL when adding data. The image will be loaded and analyzed. The resulting data is then indexed in dedicated fields.
This is great if your image collection is not huge and reindexing uncommon.
If you pre-process your raw data and store it in a database, you can pre-process your images as well and store the analysis result as JSON. When indexing your data just use the pre-processed data.
This decouples the expensive analysis process from indexing and ensures you only have to analyze an image once.
Visual searches are lightning fast due to compressed image descriptors, smart filtering, and distributed search capabilities. Searching through millions of files takes milliseconds.
Replicate your collection across server nodes. Add and remove compute capacity as needed.
Requests will be automatically balanced. If a node is down, replicas take over until the node recovers.
Split your collection across different nodes. Requests are distributed across the nodes and processed in parallel.
The HTTP REST-like API eases the integration into your tech stack. Several input and output formats are supported such as XML, JSON, CSV or binary. Many client API implementations are available for different languages including Java, PHP, Python, .NET and C#.
pixolution Flow is licensed on-premise and runs on your own servers behind your firewall. You gain full control over your search system: Choose your own hardware for optimal performance, comply with data privacy requirements and backup policies and meet internal security requirements.