Adaptive Classification of Occluded Facial Expressions of Affective States
Bibliographic record
Abstract
Internationally, the recent pandemic caused severe social changes forcing people to adopt new practices in their daily lives. One of these changes requires people to wear masks in public spaces to mitigate the spread of viral diseases. Affective state assessment (ASA) systems that rely on facial expression analysis become impaired and less effective due to the presence of visual occlusions caused by wearing masks. Therefore, ASA systems need to be future-proofed and equipped with adaptive technologies to be able to analyze and assess occluded facial expressions, particularly in the presence of masks. This paper presents an adaptive approach for classifying occluded facial expressions when human faces are partially covered with masks. We deployed an unsupervised, cosine similarity-based clustering approach exploiting the continuous nature of the extended Cohn-Kanade (CK+) dataset. The cosine similarity-based clustering resulted in twenty-one micro-expression clusters that describe minor variations of human facial expressions. Linear discriminant analysis was used to project all clusters onto lower-dimensional discriminant feature spaces, allowing for binary occlusion classification and the dynamic assessment of affective states. During the validation stage, we observed 100% accuracy when classifying faces with features extracted from the lower part of the occluded faces (occlusion detection). We observed 76.11% facial expression classification accuracy when features were gathered from the uncovered full-faces and 73.63% classification accuracy when classifying upper-facial expressions - applied when the lower part of the face is occluded. The presented system promises an improvement to visual inspection systems through an adaptive occlusion detection and facial expression classification framework.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.008 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".