HUSKYLENS 2 Face Emotion Recognition Function Description
The article offers a detailed look at HUSKYLENS 2's Face Emotion Recognition function, covering how it detects seven specific emotions, guides users through setup and operation, and provides insights into parameter adjustments for optimizing recognition performance.
1.Introduction to Face Emotion Recognition
In Face Emotion Recognition mode, HUSKYLENS 2 can recognize 7 specific expressions: Anger (ID 1), Disgust (ID 2), Fear (ID 3), Happiness(ID 4), Neutral (ID 5), Sadness (ID 6), and Surprise (ID 7). These expressions are pre-learned by HUSKYLENS 2 at the factory and do not require manual learning by the user.
2.Face Emotion Recognition User Guide
In this section, we will learn how to use HUSKYLENS 2 to recognize 7 specific human facial expressions.
2.1 Selecting the Face Emotion Recognition Function
Power on HUSKYLENS 2. After successful startup, swipe the screen to find the "Face Emotion Recognition" function.
2.2 Observe Face Emotion Recognition Results
Point HUSKYLENS 2 at an image containing faces. When a face is detected, the screen will display colored bounding boxes around all detected faces, with each face labeled above it with its expression, ID, and confidence level, such as "Happiness: ID4 98%". This means that in the Face Emotion Recognition function, the model determines there is a 98% probability that this expression corresponds to the pre-learned "Happiness" expression with ID4.
3.Face Emotion Recognition Parameter Settings
HUSKYLENS 2's factory default parameters suffice for basic functionality. If more refined control is required, individual parameters can be manually adjusted. The following parameters are specific to the "Face Emotion Recognition" function, so ensure you have entered this mode first (see the figure for reference).To modify a parameter, select it by swiping left or right on the parameter text at the bottom of the screen.
3.1 NMS Threshold
NMS Threshold is a common parameter in visual recognition used to filter detection boxes. In visual recognition tasks, models often predict multiple overlapping detection boxes around the same target. Without filtering, a single object may be boxed by multiple overlapping boxes. In such cases, adjusting the NMS threshold can remove overlapping duplicate boxes and retain only the optimal one.
In simple terms, the NMS threshold determines how much overlap between two boxes is considered "duplicate". For example, a low threshold (e.g., 0.3) will mark two boxes as duplicates if they overlap at all, removing one. A high threshold (e.g., 0.7) will only mark two boxes as duplicates if they overlap significantly, leaving more boxes in the result.
Setting Steps: Tap "NMS Threshold", and a parameter adjustment slider will appear above it. Swiping left reduces the value, and swiping right increases it.
3.2 Export Model
This parameter allows saving and exporting the currently set parameters to the local memory of HUSKYLENS 2. It is applicable to scenarios such as migrating parameters to another HUSKYLENS 2. This operation does not require inserting a TF Card.
Export Steps: Tap "Export Model". When the "Save to" pop-up window appears, slide the number up and down to select which model to save as (up to 5 models can be saved). Then tap the "Confirm" button at the bottom left of the pop-up window to save. After confirmation, the export will be automatic, as shown in the figure.
To view the exported model: After the "Exporting" pop-up disappears, you can access the exported model file on your computer.
First, connect the HUSKYLENS 2 to your computer's USB-C port.
Next, access HUSKYLENS 2's memory via your computer using the path shown in the figure below. You can view files with the extension ".json". The numbers before the file extension are the "model numbers" selected when saving the configuration. This file can be copied and pasted to other locations.
3.3 Import Model
This parameter allows importing the model exported by HUSKYLENS 2 (hereinafter referred to as "HUSKYLENS 2 A") to another HUSKYLENS 2 (hereinafter referred to as "HUSKYLENS 2 B"), so that HUSKYLENS 2 B can replicate the same parameters as HUSKYLENS 2 A without requiring re-adjustment.
Import Steps:
Step 1: Connect HUSKYLENS 2 A to the computer and copy the exported file to the computer desktop.
Step 2: Connect HUSKYLENS 2 B to the computer and paste the file from the previous step into the specified folder of HUSKYLENS 2 B, as shown in the figure. (If the "face-emotion" folder cannot be found, please complete Step 3 first. After importing the model, the "face-emotion" folder will be automatically created, and then return to Step 2.)
Step 4: Wait for the "Loading" pop-up to disappear, indicating the import is complete. Then, verify if the parameters of HUSKYLENS 2 B are consistent with those of HUSKYLENS 2 A.
Was this article helpful?
