CochleaSpecNet: An Attention Based Dual Branch Hybrid CNN-GRU Network for Speech Emotion Recognition Using Cochleagram and Spectrogram
Bibliographic record
Abstract
Being one of the main communication medium, speech contains necessary information about the emotional state of a human. Accurate emotion recognition is crucial for enhancing human-machine interactions, highlighting the importance of a strong Speech Emotion Recognition (SER) system. SER system classifies the human emotional state based on speaker’s utterances in different catagories such as sad, happy, neutral, angry, surprise, calm and so on. This research introduces a novel SER approach that utilizes cochleagram and spectrogram features to capture relevant speech patterns for the classifier network. The network integrates a hybrid model that combines Convolutional Neural Networks (CNN) for feature extraction with Gated Recurrent Units (GRU) to handle temporal dependencies. Furthermore, to improve the performance of this network, a multi-head attention mechanism has been incorporated following the GRU layer. Despite increasing interest in SER, there is a notable lack of studies using Bangla language datasets, revealing a significant gap in current research. To address this gap, evaluation of the model has been conducted on the augmented BanglaSER (Bangla Speech Emotion Recognition) dataset in which the model has achieved a notable accuracy of 92.04% in categorizing five distinct emotions: angry, surprise, happy, neutral, and sad. Additionally, to further evaluate the performance of the SER model, English language based RAVDESS (Ryerson Audio-Visual Database of Emotional Speech) dataset has also been employed into the proposed model. This attempt has provided 82.40% accuracy in classifying eight diverse emotions that includes fear, disgust, calm along with the emotions of BanglaSER. Moreover, a comparative analysis of the proposed model with existing SER approaches is carried out to demonstrate it’s stability and robustness. The incorporation of two individual features as inputs into the attention guided hybrid neural network showcases the efficacy of the proposed SER system, offering a promising approach for precise and efficient emotion categorization from speech signals.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| 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) | 0.003 | 0.001 |
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".