J. Torre 2025-12-04 NeuraCell 1.4 – Changelog This release brings substantial upgrades to supervision, detection speed, segmentation versatility, and user experience: - Segment Anything Model (SAM) Integration NeuraCell now includes Meta AI’s Segment Anything Model, enabling users to generate high-quality masks from a single click. SAM can segment virtually any structure—even in images captured from mobile phones, without the need of a custom training. - The new Playground tab allows users to test whether their images are suitable for NeuraCell’s Default model at full capability for free. Remember that if an image does not perform well, it is still possible to do a custom training. - Users can now choose their preferred mask color or use fully randomized palettes during supervision for improved visual clarity. - Added significant optimizations to reduce the response time of the assisted detection during supervision, making contextual inspections much more fluid. - The Default Mask R-CNN model has been retrained and enhanced to perform reliably on blurry, noisy, grayscale, and colorful images, improving versatility across all imaging conditions. - Masks can now be removed and recreated instantly via click, making mistake correction smoother and more intuitive. - Supported input resolutions expanded from (750–3000 px) to (10–4500 px), allowing both very high- and very low-resolution images to be processed. - FIXED: A visual issue in the Excel export where the Cell Size 2D caption appeared incorrectly placed. - FIXED: Excel export no longer draws black borders until the end of the sheet. - FIXED: A slowdown on the Supervision page when going between images. Thank you for your support and for all the ideas to improve this page. Remember that all issues can be reported at jorge.torre@uva.es.
J. Torre 2025-07-22 NeuraCell 1.3 – Changelog This update introduces major improvements to the Supervised Analysis page, offering new features and faster interaction: - Assisted Detection (Contextual Inspection): Users can now manually select a specific area of the image to inspect. This provides NeuraCell with different contexts leading to higher and more accurate detections. - Improved Mistake Removal: Multiple undesired detections can now be removed in a single action. - Supervision Finish Button: A new Finish button has been added to the supervision page, allowing users to skip ahead and directly access the final results and download options—no need to review every image. - Previous Results Revision: Users can now revisit and revise results from previous analyses directly from the Analysis page, without needing to re-upload the original images. - Download Output: When a supervised version of an image exists, the original automatic detection is no longer included in the final download, ensuring the exported results are always up to date. - Style and Performance Enhancements: The interface has been refined for a cleaner, faster user experience. Visual cues have been optimized throughout the web. - FIXED: Resolved an issue where advanced parameter characterization would go idle server-side.
J. Torre 2025-07-16 NeuraCell 1.2 – Changelog This new update brings several bug fixes, and visual improvements to the platform: - Results Table in Supervised Analysis: A new table has been added to the Supervised Analysis page, displaying the main results of the characterization directly on screen. No need to download — you can now access key data instantly. - Visual Workflow Overview: The homepage now includes a visual explanation of the NeuraCell workflow. Users can scroll through the main steps with brief descriptions and illustrative icons, improving the onboarding experience. - FIXED: Resolved an issue where the progress bar would fail if the user remained inactive or navigated away from the browser tab for an extended period. - FIXED: Corrected a bug affecting the calculation of real cell density. The reported values are now consistent with the theoretical one. Thank you again for using NeuraCell and helping us improve with your feedback!
J. Torre 2025-06-12 NeuraCell 1.1 – Changelog Several features have been added in this exciting new update: - Analysis Tab: A new Analysis section, enabling you to access and download previous results. This tab also provides access to additional post-characterization. - Post-Characterization: Added specific characterizations based on the masks generated during image inference. Remember you can request custom characterization parameters by contacting us. - Download of ROIs: You can now download the Regions of Interest (ROIs) derived from the image inference, providing you with the freedom to perform any calculations you may need. - Performance Improvements: Overall speed and robustness of the application have been significantly increased. - Security Upgrades: Internal security measures have been strengthened to ensure data protection and user privacy. - Privacy Policy Updates: Modifications have been made to the privacy policy. Check them out in https://neuracell.es/privacy_policy. Thank you so much for your support!
J. Torre 2025-04-08 NeuraCell 1.0 is Out! We are pleased to announce that NeuraCell is now available worldwide! Here is what you can do: - Analyze the structure of your porous materials using the default NeuraCell configuration. - Train on your own datasets to create a fully customized NeuraCell tailored to your specific materials. - Request additional features or parameters. Just send us an email with your needs, and we will work to integrate your custom characterization requirements. Thank you for choosing and trusting NeuraCell. Your support drives our innovation.