The ImageZebra Platform
The ImageZebra platform has been developed to track and improve the image quality of capture devices. It includes the following components:
The ImageZebra Website
Our website is the central component of the platform. At its core, it rates a scan of a test target against the FADGI General Collections specification, and allows the user to set acceptance criteria for their scans. The website tracks the performance of capture devices over time, and constructs corrections to improve device performance. With institutional partners and collaboration in mind, users who belong to a common organization have visibility of a common set of devices and results, but individualized settings and access control.
The ImageZebra Native Client
For users who can benefit from utilizing an onsite software tool to automate tracking of image quality, we provide the ImageZebra Native Client. We work with our customers to integrate the Native Client into their processes, often taking advantage of a monitored "hot folder" workflow similar to Dropbox or Google Drive. For those with more advanced needs, we offer a REST API enabling full programmatic access to our platform.
The ImageZebra ImageProcessing Technology
Based on many years of industry experience in image quality and high performance image processing, the technology built into ImageZebra results in superior corrections for optimal image quality.
- Easy drag and drop of images to analyze
- Lowest tier provides same functionality as Delt.ae
- Charting capabilities include:
- Per-check charts for Lightness, Tone Scale, DeltaE, and DeltaAB
- Geometry info (Rotation, Scale, etc)
- All image metadata
- Published API
- Each user account allows tracking of multiple devices
- Users each belong to an “organization”. Devices are owned by the organization, and can be viewed by all users of that organization with user-level access control
- Tagging devices by Customer/Project, or custom tags
- Tracks per-device performance over time
- Additional charting for: historical charts (over time) of performance for Rating, DeltaE, DeltaAB, and Tone
- Automatically sends email notification to selected recipients if a device goes in/out of spec
- ImageZebra Client Application
- Monitors an input hotfolder, waiting for images to be captured by scanner/camera
- Each image is checked to see if it is a test target or not
- Test targets are uploaded to ImageZebra
- Other images are processed with the most recent correction
- After an image is checked, it can be moved, deleted, or left in place
- Fast image processing - can utilize full hardware resources of the machine
- Runs on Mac or Windows
- Implement against other specs other than FADGI, in particular other than FADGI General collections. For example: Metamorfoze, FADGI Photographs, possibly the Imatest suite of test targets
ImageZebra Image Processing Technology
- Right now detects GoldenThread targets
- More targets can be added easily
- Images can be in any orientation
- Images can contain large margins. (Target must be at least > ½ image width)
- 2-D barcodes are decoded. Information currently not used. System could be tweaked to identify a particular device by the 2-D barcode. For example, place a 2-D barcode sticker, with the Device ID, on the test target.
- Builds a color correction using the patches from GoldenThread
- Correction is from the current state to a known colorspace, such as AdobeRGB
- Can easily build a correction from other test targets. Just need the Ideal LAB of each patch.
- Correction does NOT improve resolution. (For example, out of focus). Correction is only for color, tone and white balance.
- (MLF - Add Before/After artwork here.)
Getting Started with ImageZebra
(Last updated: 07/07/2021)
ImageZebra easily lets you check the health of your image capture system. You can get going in 3 easy steps:
Step 1: Get your system ready.
ImageZebra works by evaluating the captured images of calibrated test targets. So, in order to use the system you need to do a few things:
1) Get a physical test target
ImageZebra supports several different test targets. Use of high quality targets is an important first step in characterizing the quality of an image capture system. If you don't have a test target, one can be obtained from Image Science Associates.
2) Capture test target images
Use your image capture system (scanner or camera) to capture an image of the test target. If you don't have a test target yet, you can download a few samples here:
Step 2: Set up your ImageZebra account
Getting going with the ImageZebra platform is very easy:
1) Create an ImageZebra account.
It's easy to create an account. Once you do this, you can use ImageZebra to assess your test target images.
2) Set your preferences
The ImageZebra Settings panel is used to let the user set the desired "passing" tolerances for the various metrics. This allows the user to adjust the tolerances based on the particular circumstances of that capture device.
Step 3: Check your device
The Check Now page of ImageZebra lets you upload an image of a test target, and view the results. Uploads can be done by simply dragging and dropping an image onto "Zach", the ImageZebra, or by clicking the "Upload a file" button, and browsing to the image file.
Frequently Asked Questions
Question #1 - Why would I see differences in the results if I run a test target image on ImageZebra, and another tool such as Delt.ae?
Performing an item-to-item comparison is an ill-posed question, for the following reasons:
The name of the metrics is not consistent between the two tools. ImageZebra is using terminology as defined by FADGI.
Delt.ae does not test and report all the FADGI metrics available on the Golden Thread test pattern. Below is a list of the missing metrics not available with Delt.ae:
- OECF (Tone Response) - not measured in either colorimetric (L*) or digital counts
- White Balance Error - not measured in either colorimetric (L*) or digital counts
Resolution - FADGI guidelines have established resolution limits with star ratings. Delt.ae measures the resolution but does not report the star rating
Some metrics are only partially implemented on Delt.ae
- Noise (digital counts) - Noise measurement in units of digital counts is not available, however the Noise measurement in colorimetric units (delta L* Stdev) is available
Reproduction Scale Accuracy – Delt.ae measures the actual resolution of the scanned image, however, it does not summarize the FADGI star rating for this measurement. The summary report can be misleading when the metric is not assigned a star rating and reported with the other metrics.
Delt.ae uses the average SFR response of the 5 locations to determine the FADGI star rating. The Image Zebra and Golden Thread summary reports the worst star rating from all locations. The average location (as done by Delt.ae) could be misleading, and is not sufficiently diagnostic.
There is also a concern about the size of the analysis region used by Delt.ae, as well as differences in location where the two tools make their exact measurements. It is possible the small ROI could be giving inaccurate results.
Color Accuracy: Image Zebra currently uses the worst color patch for the summary star rating. However, this will change in the next release to be consistent with Delt.ae and Golden Thread software, which computes the Average dE color error of all color and neutral patches before determination of the star rating.
Image Zebra does not currently measure the color to color misregistration. It is unclear whether Delt.ae is using the very small ROI’s which it displays on the image. If so, there could be a large error source in the accuracy of the color registration metric. It appears Delt.ae is reporting the average color misregistration in the summary.
Delt.ae is no longer actively supported and no longer participates in any ISO standards.
Resources2016 FADGI Guidelines (pdf, 5 MB)
GoldenThread target image (tif, 16 MB)