wiktorlazarski
commented
5 months ago Hey @sweezin,
Thank you for reaching out. Sorry for such a late response.
Regarding your question:
- Estimated
offgaze_scorevalue gives an information about how likely it is that an IR image contains eye that was captured when person wasn't looking into camera, hence off gaze. Later, one can threshold that value to decide if image should be rejected or not (see this validator). - Estimated
eye_orientationvalue is an angle value expressed in degrees with respect to the horizontal screen line.
Please, let me know if that answers your question and I can close this issue.
Best regards, Wiktor
How to interpret the two output values of the offgaze_score and eye_orientation?
Context
Please provide any relevant information about your setup. This is important in case the issue is not reproducible except for under certain conditions.
Current Behavior
import cv2
img_pixels = cv2.imread("./sample_ir_image.png", cv2.IMREAD_GRAYSCALE) iris_pipeline = iris.IRISPipeline() output = iris_pipeline(img_data=img_pixels, eye_side="right") output["metadata"]
output Information (for bugs)
{'iris_version': '1.0.0', 'image_size': (1440, 1080), 'eye_side': 'right', 'eye_centers': {'iris_center': (649.6121480904019, 467.22873378530164), 'pupil_center': (664.3421158566194, 465.0919165806036)}, 'pupil_to_iris_property': {'pupil_to_iris_diameter_ratio': 0.3440108632558427, 'pupil_to_iris_center_dist_ratio': 0.05382231733026705}, 'offgaze_score': 0.00025048682955923404, 'eye_orientation': -0.0935692487493025, 'occlusion90': 0.8618111969653718, 'occlusion30': 0.9961168575085619, 'iris_bbox': {'x_min': 373.88287353515625, 'y_min': 195.31033325195312, 'x_max': 924.6512451171875, 'y_max': 738.9235229492188}}
As above, what does offgaze_score and eye_orientation mean exactly?