perClass Mira 5.4.1 Documentation

Once references are defined in the Correction tab of the Recording panel, the Recording tab is enabled. It controls the directory and filenames for scans to be stored.

The directory button specifies where the recorded scans will be stored. This directory is identical to the project Top-level data directory that can be set from the File menu or from the context menu of the Images list. The directory may be changed any time. Only the newly-recorded scans will be affected and placed in the new directory.

NOTE: perClass Mira never deletes or moves any scans. It is user responsibility to perform any destructive actions on the data, if needed.

The scan field defines the name of the scan to be recorded. Note the bigger font used for the scan name. This is because perClass Mira Stage users can fully operate the scanning process using only the stage buttons at a larger distance from the computer and the screen.

NOTE: The scan name must be defined in order to record a scan. When working with the perClass Mira Stage, issuing a Scan and Record command will not work unless the scan name is defined.

Making scan names unique upon saving

The Record button starts the data recording in Camera belt or waterfall modes. In the scan mode, this button invokes saving of the data. In any of these operations, we may wish to make the scan name unique automatically. We have two options:

• Appending a time-stamp.
• Appending an index (to be precise an underscore, followed by an index), the index auto-increments after each recording.

TIP: When recording scans with training samples, it is recommended to use only samples of one class and use its name as a scan name.

Manual incrementing of scan indices

If we wish to increment indices in scan names manually, perClass Mira provides us very simple but efficient assistance. If a scan name is a set of letters followed by a number, it is recognized as an "incrementable" pattern and it is colored in green. Compare the following image with the screenshot above where the scan name did not end with a digit.

The user may increment the index using:

• Cursor up / down keys
• Mouse wheel
• Stage button assigned to Increment scan name command

Meta-data information in scan names

In larger projects, it is recommended to store different types of meta-data information about the samples in the scan filenames. It is, for example, useful to distinguish the day of scanning, supplier, fruit variety, and others. The main advantage is that we may easily select images in our project based on these textual patterns. This helps us to test or cross-validate our models on unseen days, suppliers, or varieties.

The scan name widget recognizes underscores as boundary between name items. The currently selected item is rendered in green, other items are colored in red. We may select the item by clicking on it or positioning the cursor. 

The listed methods of incrementing the scan name mentioned above then apply to the current item.

Scan name templates

perClass Mira 5.2.1 added support for scan names defined by a template:

The template field below the scan name is used to define different sections, separated by an underscore, which have different formats defining how they will appear in the scan name and how they will increment. Focusing on this field will show the following popup below it to give a short explanation about the available options:

• Numbers: sections using this format will have incrementing numbers with a minimum width of the amount of #s used.
• obj### will become obj001, incrementing will make it obj002, obj003, obj004, ..., obj999, obj1000, obj1001.
• Range: sections using this format will have incrementing numbers within a specified range, optionally with a step number, values will wrap around.
• rep{1:3} will become rep1, incrementing will make it rep2, rep3, rep1.
• obj{1:2:10} will become obj1, incrementing will make it obj3, obj5, obj7, obj9, obj1.
• Enumeration: sections using this format will iterate over the different values provided, values will wrap around.
• color{R|G|B} will become colorR, incrementing will make it colorG, colorB, colorR.
• Date time: sections using this format will be formatted using the current date and time. Note that these sections cannot be incremented, instead they will be formatted when starting.
• %FT%T will become 2025-04-24T13:45:00.

The template will be saved to the mira.ini file, as well as the current values, this way the scanning can continue easily on the next day.

NOTE: When using a template, it will not be possible to modify the scan name manually, only the sections can be incremented.

NOTE: Modifying the template will reset the values back to their initial values.

A selected session can be incremented using the Up arrow key or scroll wheel up and decremented using the Down arrow key or scroll wheel down. By default the section will increment (or decrement) by 1 step, if the Control key is held, then it will increment by 10 steps instead.

Number sections can also be modified by double clicking the section and typing the number. A number shorter than the minimum width will automatically be padded.

Date time format specifiers

Below is a list of commonly used options for date time formatting, see the documentation for strftime for all options. Note that %n and %t are not supported.