> For the complete documentation index, see [llms.txt](https://docs.biodock.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.biodock.ai/analysis-and-results/correcting-results-to-improve-your-model.md).

# Correcting results to improve your model

*What you'll get: how to send an image from a result back to your project as labels, correct only what the model got wrong, and train a better version.*

New images often differ from the ones you trained on: a new cell line, stain or microscope. The model may then miss objects or add wrong ones. Instead of labeling those images from scratch, import the model's predictions into your project as labels, fix the mistakes, and train again.

{% hint style="info" %}
**Edit objects or Import to project?** [Editing objects](https://docs.biodock.ai/analysis-and-results/editing-objects-qc.md) fixes this one result. Importing to a project fixes the model, so future results improve. You can do both.
{% endhint %}

## Import an image's predictions

You must own the target project. Import works for segmentation results, and not in the full-screen viewer. Importing does not charge storage.

<figure><img src="https://3806122466-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk3DfNQKoMy1JXdaDLvpG%2Fuploads%2Fgit-blob-76346370ff0a03985da4666536b8dc6531b82b38%2Fdocs-results-import-to-project.gif?alt=media" alt="Under the image viewer open the More actions menu, click Import to project, keep Nuclei counter as the Target project, map Nucleus to Nucleus, click Continue, tick tiles, click Continue, click Import"><figcaption><p><strong>⋯</strong> → <strong>Import to project</strong> → map the classes → pick tiles → <strong>Import</strong>.</p></figcaption></figure>

1. On the result's **Dashboard** tab, go to the image you want to correct in the image viewer.
2. Under the viewer, open the **⋯** (More actions) menu and choose **Import to project**. This imports the image you are viewing.
3. In the **Add predictions to a project** window, choose the **Target project**. Only projects you own whose images have the same number of channels as these results are listed. The project the model came from is preselected if you are a member of it.
4. Map each **Current result class** to a **Target project class**. Pick an existing class, or **Add new class**. If you have renamed classes, check this mapping.
5. Click **Continue**. If asked, select up to 10 tiles of the image for each size group, and click **Continue** again.
6. Check the summary and click **Import**.

The window closes while the import runs. When it is done, an **Image added to project** notification arrives in the bell in the top bar, with a **Go to project** link.

## Correct and retrain

1. In the project, open the imported image on the labeling screen.
2. Correct the imported tiles, focusing on where the model did badly. Delete wrong objects, add missed ones and fix rough outlines with the [labeling tools](https://docs.biodock.ai/projects-and-models/labeling-tools.md). Make sure every object in the tile is labeled correctly.
3. Hover over each corrected tile and tick **Fully labeled** in its corner. Only fully labeled tiles are used for training.

<figure><img src="https://3806122466-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk3DfNQKoMy1JXdaDLvpG%2Fuploads%2Fgit-blob-1f3f03a7d71e693fb62527a95fbc1c4da6d47883%2Fdocs-labeling-fully-labeled-tile.png?alt=media" alt="A tile corner with the label count and the Fully labeled tick box ticked, and a green tile border"><figcaption><p>Tick <strong>Fully labeled</strong> once a tile is fully corrected.</p></figcaption></figure>

Once you have corrected tiles from a few images, [train a new version](https://docs.biodock.ai/projects-and-models/training-a-model.md). It should do much better on images like the ones you corrected. Compare it with the earlier version in [Evaluating performance](https://docs.biodock.ai/projects-and-models/evaluating-performance.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.biodock.ai/analysis-and-results/correcting-results-to-improve-your-model.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
