View on GitHub

Object Finder

3D object recognition in volumetric images

Download Object Finder Release notes User manual Report a bug Native builds:

What is Object Finder?

Object Finder is a MATLABĀ® app that allows you to recognize small structures in a large volumetric image, plot their statistics and perform quantitative analysis such as nearest neighbor and colocalization analysis.

This software is developed for neuroscience research purposes, where it is able to detect fluorescently-labeled synapses in neuronal image stacks acquired using confocal or super-resolution microscopes.

Object Finder is licensed under GNU General Public License v3. This strong copyleft license ensures you will always have the right to obtain, view and modify Object Finder's source code to review its functionality and improve its features.

We strongly believe that software developed for data analysis in scientific research must be open source, to ensure the highest level of reproducibility of your science.

Whether your workstation relies on Microsoft Windows, macOS or Linux, Object Finder can run on your computer.

Visually interact with your data

Object Finder can display detected objects using the builtin volume inspector on top of your source image volume. Explore, filter and interact directly with detected objects.

You can visually add new objects or refine existing ones, visually inspect them against the raw signal and isolate subsets of objects based on their morphology.

Need to render your objects? Object Finder seamlessly supports ImarisĀ® to take full advantage of Imaris advanced 3D rendering capabilities.

Plot results and export data to Excel

Object Finder calculates distribution of objects properties in the volume such as: density along Z or along a skeleton, location, brightness and shape of objects. Analysis results can be exported to spreadsheets for further elaboration or plotting using your favorite data analysis application.

Colocalization analysis

Manually colocalize objects with co-labeled structures or let Object Finder perform automatic colocalization analysis between sets of objects or against a binary mask.

Deep Learning at your service

Do you have millions of candidate to validate as objects vs noise? Train your personal deep learning network using you own trusted data to automatize and reduce subjectivity of this critical step.

Object Finder uses the latest deep learning technology together with GPU computing to scale up the validation of large number of objects. This feature takes advantage of recent NVIDIA graphic cards through the Neural Network toolbox.

Works around your schedule

Do you have many images to process? Use the automation tab to allow batch processing without user supervision.

Schedule Object Finder to perform repetitive task automatically during your quiet hours, so you can keep working on your other projects.

How do I install Object Finder?

  1. Download latest Object Finder version from the top of this page
  2. Install Object Finder using MATLAB's "Install App" button
  3. Start Object Finder from your Apps list.
  4. Head to the user manual to learn about all Object Finder functionalities

System Requirements

Software: Hardware (recommended):
  • CPU: Intel Xeon >2GHz
  • RAM: 32Gb
  • GPU: NVidia for deep learning