/** * SPDX-FileCopyrightText: Copyright contributors to the CopilotJ project. * * SPDX-License-Identifier: Apache-2.0 */ export type Tag = | "All" | "ImageJ/Fiji" | "Python DL" | "Segmentation" | "3D/Visualization" | "Tracking" | "Spatial-omics" | "Research/Web"; export const GITHUB_REPO = "https://github.com/neurogeom/copilotj"; export const DISCORD_LINK = "https://discord.gg/32wpPq5dBV"; export const IMAGESC_LINK = "https://forum.image.sc/tag/copilotj"; export const SLOGAN = "CopilotJ turns natural-language requests into executable, verifiable bioimage analysis workflows by coordinating ImageJ/Fiji ecosystem, Python scientific libraries and deep-learning tools"; export const FEATURES = [ { label: "Intelligent", description: "Users only need to describe analysis tasks in conversation, and CopilotJ iteratively generates structured, actionable steps to complete end-to-end analyses", }, { label: "Efficient", description: "Compared with external chatbot–assisted ImageJ workflows, CopilotJ improves user efficiency and reduces interaction burden, while substantially enhancing analytical capability and narrowing the performance gap between beginners and experienced users", }, { label: "Reproducible", description: "Successful runs can be exported as reusable workflows and replayed to reproduce results and scale analyses across datasets with minimal user intervention", }, ]; export const BACKGROUND = [ "Quantitative bioimage analysis is indispensable to modern life science research, but ImageJ workflows remain difficult to scale beyond experience-reliant, menu-driven interaction and macro scripting", "The ImageJ and Fiji ecosystem offers rich plugin coverage, yet researchers still face fragmented documentation, parameter-heavy tools, and limited access to modern Python and deep-learning workflows", "CopilotJ addresses this gap with an LLM-based conversational interface that maps high-level intent to executable ImageJ-centric analysis procedures.", ]; export const CONTRIBUTIONS = [ { title: "Human-in-the-loop (HITL)", description: "CopilotJ supports human-in-the-loop interaction, allowing users to intervene at critical decision points through familiar GUI operations while maintaining workflow continuity", }, { title: "Vision-assisted self-reflection", description: "CopilotJ integrates feedback from tool outcomes, software states, and visual observations to perform reflective reasoning and generate corrective scripts or revised procedures when needed", }, { title: "Self-evolving", description: "CopilotJ maintains a self-evolving knowledge bank that captures and serializes successful workflows, reusing prior experience to guide planning and tool selection for new tasks", }, ]; export const SUPPLEMENTARY_VIDEOS = [ { id: 1, group: "Foundational operations", title: "Basic manipulation", youtubeUrl: "https://www.youtube.com/watch?v=QdeXvtg9u_o", summary: "Executes a multi-step ImageJ workflow including duplication, rotation, contrast adjustment, blur, Otsu thresholding, and edge detection", }, { id: 2, group: "Foundational operations", title: "Video generation", youtubeUrl: "https://www.youtube.com/watch?v=9r3M0G6EI0g", summary: "Generates a 3D visualization video from neuron image data with automatic denoising and contrast enhancement when visibility is poor", }, { id: 3, group: "Foundational operations", title: "Batch post-processing", youtubeUrl: "https://www.youtube.com/watch?v=JLcoPdPfYEI", summary: "Processes folders of microscopy images in batch by adjusting contrast, adding scale bars, and exporting standardized outputs", }, { id: 4, group: "Core computational imaging", title: "3D deconvolution", youtubeUrl: "https://www.youtube.com/watch?v=Nc6xSoQmJWI", summary: "Performs image deconvolution on a Drosophila stack using a provided PSF and exports the restored volume for analysis", }, { id: 5, group: "Core computational imaging", title: "Human-in-the-loop segmentation", youtubeUrl: "https://www.youtube.com/watch?v=3yVhDykL6fs", summary: "Surveys candidate segmentation strategies, lets the user refine the threshold inside ImageJ, and produces the final segmentation under transparent HITL control", }, { id: 6, group: "Core computational imaging", title: "Object tracking", youtubeUrl: "https://www.youtube.com/watch?v=6DNCgnTNy48", summary: "Detects and tracks objects in time-lapse microscopy movies and exports motion statistics together with trajectory visualizations", }, { id: 7, group: "Domain-specific quantification", title: "Colocalization analysis", youtubeUrl: "https://www.youtube.com/watch?v=r1N9ctI-5qA", summary: "Runs red-green channel colocalization analysis and exports standardized metrics plus analysis outputs", }, { id: 8, group: "Domain-specific quantification", title: "DNA comet assay quantification", youtubeUrl: "https://www.youtube.com/watch?v=RyVIedlpw6k", summary: "Segments comet heads and tails, quantifies tail length and DNA percentage in the tail, and writes overlays and measurements", }, { id: 9, group: "Domain-specific quantification", title: "Gel electrophoresis analysis", youtubeUrl: "https://www.youtube.com/watch?v=Y88URkSEi8Y", summary: "Quantifies lane-wise signal from SDS-PAGE imagery and automatically summarizes results as an interpretable bar chart", }, { id: 10, group: "Domain-specific quantification", title: "Cell scratch assay measurement", youtubeUrl: "https://www.youtube.com/watch?v=YHg8DLboKdw", summary: "Measures wound area at multiple time points and computes wound-closure percentage with a maximum-gap strategy", }, { id: 11, group: "Multi-stage pipelines", title: "Glomerulus quantification and workflow management", youtubeUrl: "https://www.youtube.com/watch?v=CY-3c3HCMjo", summary: "Builds a hybrid ImageJ-plus-Python pipeline for tissue morphology analysis, then saves and replays it as a reusable batch workflow", }, { id: 12, group: "Multi-stage pipelines", title: "Automated calcium imaging analysis", youtubeUrl: "https://www.youtube.com/watch?v=q53SQrylGpE", summary: "Runs motion correction, neuron segmentation, ROI extraction, response alignment, response classification, and summary export for calcium imaging stacks", }, { id: 13, group: "Benchmarking", title: "Comparison with interactively used ImageJ", youtubeUrl: "https://www.youtube.com/watch?v=7l5iEjIt3rg", summary: "Contrasts CopilotJ’s prompt-driven automation with manually executed ImageJ operations in a side-by-side workflow comparison", }, { id: 14, group: "Deep-learning extensions", title: "Image super-resolution", youtubeUrl: "https://www.youtube.com/watch?v=kjzNdAof_oY", summary: "Invokes a BiaPy-based super-resolution model to enhance image clarity and structural detail for downstream analysis", }, { id: 15, group: "Deep-learning extensions", title: "Cell classification", youtubeUrl: "https://www.youtube.com/watch?v=sCf8_dEHC1Q", summary: "Runs pretrained deep-learning inference to classify cell types and returns structured prediction outputs", }, { id: 16, group: "Deep-learning extensions", title: "Cell profiling with Cellpose", youtubeUrl: "https://www.youtube.com/watch?v=nYcLd9dMnow", summary: "Combines Cellpose segmentation with classical processing steps to generate quantitative cell profiling outputs", }, ] as const; export const ORGS = [ { name: "Guangdong Institute of Intelligence Science and Technology", logo: "/imgs/logo_gdiist.png", link: "https://www.gdiist.cn/", }, { name: "The Hong Kong Polytechnic University", logo: "/imgs/logo_polyu.png", link: "https://www.polyu.edu.hk/" }, { name: "Beijing Normal University", logo: "/imgs/logo_bnu.png", link: "https://www.bnu.edu.cn/" }, ]; export const TOOLS = [ // Root directory tools { name: "Bio-Formats", logo: new URL("./assets/tools/logo_bioformats.png", import.meta.url).href, link: "https://www.openmicroscopy.org/bio-formats/", }, { name: "BioImage.IO", logo: new URL("./assets/tools/logo_bioimage-io-icon.svg", import.meta.url).href, link: "https://bioimage.io/", }, { name: "BoneJ", logo: new URL("./assets/tools/logo_bonej.png", import.meta.url).href, link: "https://bonej.org/" }, { name: "Cellpose", logo: new URL("./assets/tools/logo_cellpose.png", import.meta.url).href, link: "https://www.cellpose.org/", }, { name: "CLIJ2", logo: new URL("./assets/tools/clij2_logo.png", import.meta.url).href, link: "https://clij.github.io/", }, { name: "CSBDeep", logo: new URL("./assets/tools/logo_csbdeep.png", import.meta.url).href, link: "https://csbdeep.bioimagecomputing.com/", }, { name: "DeepImageJ", logo: new URL("./assets/tools/logo_deepimagej.png", import.meta.url).href, link: "https://deepimagej.github.io/", }, { name: "FLIMLib", logo: new URL("./assets/tools/logo_flimlib.png", import.meta.url).href, link: "https://flimlib.github.io/", }, { name: "Image.sc Forum", logo: new URL("./assets/tools/logo_imagesc.png", import.meta.url).href, link: "https://forum.image.sc/", }, { name: "ImgLib2", logo: new URL("./assets/tools/logo_imglib2.png", import.meta.url).href, link: "https://imglib2.net/", }, { name: "OMERO", logo: new URL("./assets/tools/logo_omero.png", import.meta.url).href, link: "https://www.openmicroscopy.org/omero/", }, { name: "StarDist", logo: new URL("./assets/tools/logo_stardist.png", import.meta.url).href, link: "https://github.com/stardist/stardist", }, { name: "TrackMate", logo: new URL("./assets/tools/logo_trackmate.png", import.meta.url).href, link: "https://imagej.net/plugins/trackmate/", }, // Python tools { name: "BiaPy", logo: new URL("./assets/tools/python/logo_biapy.png", import.meta.url).href, link: "https://biapy.readthedocs.io/", }, { name: "Python", logo: new URL("./assets/tools/python/logo_python.png", import.meta.url).href, link: "https://www.python.org/", }, { name: "PyTorch", logo: new URL("./assets/tools/python/logo_pytorch.svg", import.meta.url).href, link: "https://pytorch.org/", }, { name: "TensorFlow", logo: new URL("./assets/tools/python/logo_tensorflow.svg", import.meta.url).href, link: "https://www.tensorflow.org/", }, { name: "OpenCV", logo: new URL("./assets/tools/python/logo_opencv.png", import.meta.url).href, link: "https://opencv.org/", }, // Research tools { name: "DuckDuckGo", logo: new URL("./assets/tools/research/logo_ddg.svg", import.meta.url).href, link: "https://duckduckgo.com/", }, { name: "Google", logo: new URL("./assets/tools/research/logo_google.png", import.meta.url).href, link: "https://www.google.com/", }, { name: "Tavily", logo: new URL("./assets/tools/research/logo_tavily.svg", import.meta.url).href, link: "https://tavily.com/", }, { name: "Wikipedia", logo: new URL("./assets/tools/research/logo_wikipedia.png", import.meta.url).href, link: "https://www.wikipedia.org/", }, // // LLM tools // { name: "DeepSeek", logo: new URL("./assets/tools/llms/logo_deepseek.png", import.meta.url).href, link: "https://www.deepseek.com/" }, // { name: "Gemini", logo: new URL("./assets/tools/llms/logo_gemini.jpg", import.meta.url).href, link: "https://gemini.google.com/" }, // { name: "GLM", logo: new URL("./assets/tools/llms/logo_glm.svg", import.meta.url).href, link: "https://zhipuai.cn/" }, // { name: "GPT", logo: new URL("./assets/tools/llms/logo_gpt.png", import.meta.url).href, link: "https://openai.com/" }, // { name: "Ollama", logo: new URL("./assets/tools/llms/logo_ollama.webp", import.meta.url).href, link: "https://ollama.com/" }, ];