User Interface

Startup Interface

When you start the Mech-DLK software, the following startup interface appears:

New Project

Create a new empty project.

Open Project

Select and open an existing project.

Example Projects

View an example project.

More Projects

Download more example projects for use.

Go to the startup interface.

Home Interface

When you create or open a project, the following home interface appears. You can perform operations such as data viewing and labeling, model training and validation.

This interface consists of the following parts:

menu bar

Includes File, Edit, Tools, Settings, and Help.

project tab bar

Displays all open projects.

image list

Includes image processing tools, the Import/Export button, labeling tools, and the image list.

labeling toolbar

Includes the tools that are used to label images.

labeling area

Displays and labels the current image.

module area

Includes all modules and function buttons in the current project.

display settings

Modifies the display settings.

feature tabs

Includes the Labeling tab, Training tab, Validation tab, and function buttons.

Menu	Option	Description
File	New Project	Create a new project.
Open Project	Select and open an existing project folder.
Save Project	Save changes to the current project.
Save As	Save the project under a specified directory.
Edit	Undo	Undo the Labeling action.
Redo	Redo the action that you’ve undone.
Tool	Operation mode	You can perform batch-wise testing on the dataset under the Operation Mode. This feature is for the Defect Segmentation module and the Classification module.
Training Center	Manage training tasks and training history.
Settings	Settings	Change the display language, turn on/off the Developer Mode, and set whether to always load the smart labeling model.
Help	User Manual	View the user manual.
Quick Guide	Quick guide to start the software.
What’s New	View the release notes.
Logs	Open the folder containing the running logs.
Device Check	Check and view the CPU and GPU information of the deployment device.
About	View the software version.

Option

Description

File

New Project

Create a new project.

Open Project

Select and open an existing project folder.

Save Project

Save changes to the current project.

Save As

Save the project under a specified directory.

Edit

Undo

Undo the Labeling action.

Redo

Redo the action that you’ve undone.

Tool

Operation mode

You can perform batch-wise testing on the dataset under the Operation Mode.
This feature is for the Defect Segmentation module and the Classification module.

Training Center

Manage training tasks and training history.

Settings

Change the display language, turn on/off the Developer Mode, and set whether to always load the smart labeling model.

Help

User Manual

View the user manual.

Quick Guide

Quick guide to start the software.

What’s New

View the release notes.

Logs

Open the folder containing the running logs.

Device Check

Check and view the CPU and GPU information of the deployment device.

About

View the software version.

Project Tab Bar

This bar displays all open projects. You can click the project tabs to switch between projects.

Image List

Feature	Option	Description
search box	Search for images in the image list.
image processing tools	Preprocessing Image	Adjust the brightness, contrast, and color balance of images.
Filter Images	Set the filtering conditions to conveniently select specific images from datasets.
Switch to Preview Mode	Switch the image display mode to preview view or list view.
Partition Training and Validation Sets	Drag the slider to adjust the proportion of the training set and validation set.
Import/Export	Import from Previous Module	Import data from the previous module (applicable to cascaded modules).
Import Image	Import one or more images.
Import Folder	Import all images from the selected folder, excluding images in subfolders.
Import Dataset	A dataset contains images and corresponding labels. * Support the import of datasets in the DLKDB format (.dlkdb), which are datasets exported from Mech-DLK. * Support the import of datasets in the COCO format (applicable to the Instance Segmentation and Object Detection modules).
Export Dataset	Export the datasets from the current project.
labeling tools	Pre-trained labeling	Use the Pre-trained Labeling Tool to label all images in the image list.
VFM labeling	Use the VFM Labeling Tool to label all images in the image list.
image list	Displays all images and their attributes in the current project. You can select one or more images, then right-click the images to perform corresponding operations.

Feature

Option

Description

search box

Search for images in the image list.

image processing tools

Preprocessing Image

Adjust the brightness, contrast, and color balance of images.

Filter Images

Set the filtering conditions to conveniently select specific images from datasets.

Switch to Preview Mode

Switch the image display mode to preview view or list view.

Partition Training and Validation Sets

Drag the slider to adjust the proportion of the training set and validation set.

Import/Export

Import from Previous Module

Import data from the previous module (applicable to cascaded modules).

Import Image

Import one or more images.

Import Folder

Import all images from the selected folder, excluding images in subfolders.

Import Dataset

A dataset contains images and corresponding labels.
* Support the import of datasets in the DLKDB format (.dlkdb), which are datasets exported from Mech-DLK.
* Support the import of datasets in the COCO format (applicable to the Instance Segmentation and Object Detection modules).

Export Dataset

Export the datasets from the current project.

labeling tools

Pre-trained labeling

Use the Pre-trained Labeling Tool to label all images in the image list.

VFM labeling

Use the VFM Labeling Tool to label all images in the image list.

image list

Displays all images and their attributes in the current project. You can select one or more images, then right-click the images to perform corresponding operations.

Labeling Toolbar

Mech-DLK provides various image labeling tools to label the objects in the images according to your needs. You can view the supported labeling tools and their usage details in each algorithm module.

Category	Tool	Supported Module
label tool	OK Label	Defect Segmentation, Unsupervised Segmentation
NG Label	Unsupervised Segmentation
labeling tool	Polygon Tool	Instance Segmentation, Defect Segmentation, Object Detection, Text Detection, Fast Positioning
Ellipse Tool	Instance Segmentation, Object Detection, Fast Positioning
Rectangle Tool	Instance Segmentation, Object Detection, Fast Positioning
Free Rectangle Tool	Text Detection
Brush Tool	Defect Segmentation
Autofill Lasso Tool	Defect Segmentation
Smart Labeling Tool	Instance Segmentation, Defect Segmentation, Object Detection, Fast Positioning
Pre-trained Labeling Tool	Instance Segmentation, Defect Segmentation, Classification, Object Detection, Text Detection
VFM Labeling Tool	Instance Segmentation, Object Detection, Text Detection
mask tool	Mask Polygon Tool	Instance Segmentation, Unsupervised Segmentation, Defect Segmentation, Object Detection, Fast Positioning
Mask Brush Tool	Instance Segmentation, Unsupervised Segmentation, Defect Segmentation, Object Detection, Fast Positioning
Mask Lasso Tool	Instance Segmentation, Unsupervised Segmentation, Defect Segmentation, Object Detection, Fast Positioning
eraser tool	Labeling Eraser Tool	Defect Segmentation
Mask Eraser Tool	Instance Segmentation, Unsupervised Segmentation, Defect Segmentation, Object Detection, Fast Positioning
grid tool	Grid Cutting Tool	Defect Segmentation
Grid Selection Tool	Defect Segmentation
Text Recognition Tool	Text Recognition
Template Tool	Instance Segmentation, Object Detection, Text Detection
ROI Tool	Instance Segmentation, Defect Segmentation, Classification, Object Detection, Fast Positioning, Text Detection, Text Recognition, Unsupervised Segmentation
Select Tool	Instance Segmentation, Defect Segmentation, Classification, Object Detection, Fast Positioning, Text Detection, Text Recognition, Unsupervised Segmentation

Labeling Area

This area displays the current image, labels on the image, and attributes. You can perform the following operations in this area:

Select labeling tools and label the image.
Right-click on the image and labels to perform the corresponding operations.
Scroll upwards or downwards to zoom in or out the image.
Move the cursor to the left or right sides of the image to quickly switch between previous and next images.
View the keyboard shortcuts.
For some algorithm modules, the image label set by the user is displayed in the upper left corner of the image. After validation, the upper right corner displays the label results, hardware type, and validation duration.

Module Area

This area displays a single module or multiple modules of the project. You can perform the following operations in this area:

Click the Add button to add a new module for a new project.
Click the Add button to add a cascaded module for the existing module.
Click the Delete button to delete a module.

Display Settings

You can show or hide views and adjust view opacity in this area.

Feature Tabs

This area includes the Labeling Tab, the Training Tab, the Validation Tab, and function buttons.

Labeling Tab

The Labeling Tab displays the label classes and image tags in the current project. You can perform the following operations in the Labeling Tab:

In the Classes area,
- Create or delete the classes.
- After selecting a class, right-click the class to edit, delete, or merge it.
In the Image Tag area,
- Create or delete tags.
- Set a tag for the current image.
- Double-click the tag name to rename the tag.

Training Tab

In the Training Tab, you can configure training parameters and view training information.

For more information about detailed parameters and training guidance, see Train a model.

Validation Tab

After model training, you can configure validation parameters and view the validation results of models in the Validation Tab.

For more information about detailed parameters and validation guide, see Validate a model.

Function Buttons

In the Training Tab and Validation Tab, you can perform operations by clicking the Train, Validate, and Export buttons.

Mech-DLK provides corresponding function buttons for algorithm modules. For more information, see the introductions of each algorithm module.