Skip to main content
AboutGuideContact
Sign in / Register

Analyze Images

The Analysis workspace is the hub for image-based research. Import microscopy images, run guided workflows that walk you through detection step by step, train a custom AI model on your own labels, and pull quantitative results into tables and figures.

Placeholder: Screenshot Coming Soon

Analysis workspace gallery showing microscopy thumbnails and the category tabs at the top

The Analysis Hub

The Analysis sidebar entry expands into a set of category tabs:

  • Imaging for microscopy image analysis (this page)
  • Flow for flow cytometry data (see Analyze flow cytometry)
  • Spectra for spectroscopy (placeholder, not yet built out)
  • Chromato for chromatography (placeholder, not yet built out)

Imaging and Flow are the working surfaces today.

Image Gallery

The gallery is where you browse and organize the images attached to a project. Supported upload formats are TIFF, PNG, JPEG, and LIF (Leica's microscopy format). Large files stream as they upload, so multi-gigabyte LIF stacks don't choke your browser. LIF files with multiple series let you pick which series to open.

Importing Images

Click Upload to add images from your computer. The upload dialog box accepts the formats above. Multi-series LIF files prompt you to select which series to bring in.

Note on LIF files: Conspecta extracts each series from a LIF upload as a standard image (TIFF, PNG, or JPEG) and keeps those instead of the original LIF. The extracted images are what you actually view, annotate, and analyze — keeping the original LIF around would just mean every viewer has to re-parse a multi-series microscopy container before showing you a picture. If you need the original LIF for archival, keep your own copy.

Z-Stacks in the Gallery

Multi-plane Z-stack images display a compact badge in the bottom-left corner of each thumbnail. The badge shows the slice and channel counts separated by a dot, for example 10z · 3ch (ten Z-planes, three fluorescence channels). If an image has only one plane or one channel, only the relevant number appears. In table view the same information is split into separate Slices and Channels columns so you can sort by either.

Organizing

  • Folders to group images by experiment, date, or sample
  • Stars to pin the images you're actively working on
  • Bulk selection to move, delete, or star many at once

Views

  • Grid with thumbnails for visual recognition (the default)
  • Table for sortable, filterable metadata views

Analysis Workflows

When you open an image to analyze, the workspace offers a set of guided workflows, each tuned to a common research question. The workflow walks you through the right tools in the right order, with previews that update in real time as you adjust parameters.

Live Workflows

  • Object Count for counting nuclei, colonies, puncta, cells, or any discrete object population
  • Confluency / Area Coverage for measuring confluency, wound healing area, staining coverage, or any area-based metric
  • Marker Intensity for quantifying fluorescence intensity or signal strength per object
  • Marker Classification for classifying detected objects as positive or negative based on a marker channel
  • Colocalization for measuring how strongly two channels co-localize, with Pearson r, Manders M1/M2, and automatic Costes thresholding

Workflow Steps

Most workflows follow the same pattern:

  1. Isolate to detect objects with thresholding
  2. Separate to split touching or overlapping objects
  3. Filter to drop objects that are too small, too large, or the wrong shape
  4. Run to execute the analysis and write results to the Data tab

Area-based workflows (like Confluency) skip the separate and filter steps and go straight from threshold to result.

Each step opens a configuration popover. As you nudge the threshold, watch the preview update on the image. Run only when you're happy with the preview.

Tip: Workflows aren't one-shot. You can adjust an earlier step and re-run; the workspace recomputes the downstream steps without losing your selections.

Narrow your count: In Object Count you can filter the detected objects by size or color right in the results, so a count targets just the population you care about, like the small nuclei or the red-stained cells.

Counting a Z-stack: For a multi-plane image, the What to count panel lets you choose what each image counts. The default is the maximum intensity projection, which counts the whole stack at once so a cell that spans several planes is counted once rather than on every plane. You can instead count a single slice. Each image remembers its choice, and "Count all" applies each image's own setting. After counting, the results plot's Show menu adds a By depth view for a stack — use Break down by depth to count every plane and see how objects are distributed through it, with the busiest plane marked.

Custom AI Models

For objects that don't fit a threshold-and-filter workflow, train a Custom AI Model on your own examples (or run a pre-trained model) to detect them directly. Start one from scratch or seed it from an Object Count run. See Image AI and data for the full guide.

Next Steps

Work with suppliersImage tools