Plainsight Pipelines is the latest addition to our end-to-end vision AI platform and a foundational technology that supports our custom enterprise solution development. This exciting new feature makes it easy to see vision AI success by making visual data workflows repeatable, scalable, and traceable.
Centralize and streamline visual data workflows: Pipelines supports both streaming inputs for users who need always-on pipelines and batch inputs for making scheduled uploads.
Run custom created computer vision models: Train custom models within the Plainsight vision AI platform and bring your models into a pipeline to setup custom prediction pipelines.
Apply uniform visual data augmentations at scale: Processing blocks like image resizing, cropping, and tiling provide repeatable and uniform application of visual data augmentations.
Visual data workflow uniformity is simple with Plainsight Pipelines. Here’s how to get started today.
Step 1: Access Plainsight Pipelines
Register for the Plainsight vision AI platform or log into your existing account to get started.
Step 2: Select the Appropriate Type of Pipeline and Name It
With Plainsight Pipelines, you can build, manage, and deploy both batch and streaming pipelines.
Step 3: Connect Inputs to Your Pipeline
Streamline data uploads by connecting your pipeline to Google Cloud Storage, Pub/Sub, API Endpoints or Amazon S3 buckets.
Step 4: Select a Prediction Block
Choose an off-the-shelf model like logo or landmark detection or select a custom trained computer vision model.
Step 5: Apply Transformation Blocks
Next, apply transformation blocks to crop or resize your visual data.
Step 6: Define Your Outputs and Publish Your Pipeline
Finally, once you’ve selected the type of output you’re looking for, it’s time to publish your pipeline.
Learn More and Share Your Thoughts
We are excited to launch this new feature to both On-Demand and Enterprise customers and look forward to hearing how Plainsight Pipelines is empowering your business and driving repeatable, traceable and scalable computer vision innovation.