ARNet: Enhanced Tracking and Optimization for Video-Based Affect Recognition
Bibliographic record
Abstract
Affect recognition plays a vital role in understanding individuals' emotional well-being and social interactions, particularly for children with autism who often face challenges in expressing their emotions.In this paper, we propose a novel approach to enhance affect recognition in children with autism by combining the power of deep learning with a new optimization method called "Stochastic Average Gradient Augmented with Tracking" (SAGAT).Through extensive experimentation, we demonstrate that the proposed approach significantly improves the accuracy of the Convolutional Neural Network (CNN) model compared to conventional optimization methods.Notably, our approach demonstrated strong accuracy in predicting arousal and valence in children with autism, with the CNN model achieving a mean squared error of 0.225 for arousal and 0.174 for valence on the SSBD-affect dataset.Pre-training on the AffectNet dataset further improved performance, reducing MSE to 0.187 for arousal and 0.156 for valence, highlighting the benefits of transfer learning.These findings hold significant implications for facilitating better understanding, support, and intervention strategies for children with autism, ultimately fostering their emotional well-being and social integration.This research opens up promising avenues for integrating advanced optimization techniques with deep learning to empower individuals with autism and promote inclusive technologies in affective computing.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.002 |
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 source (direct Gemma or distilled Codex), 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".