•  
  •  
 

Document Type

Original Article

Abstract

In recent studies, computer vision and human-computer interaction have focused heavily on facial expression recognition (FER). Traditional deep learning algorithms have faced significant challenges when processing images with occlusion, uneven lighting, and positional inconsistencies and addressing dataset imbalances. These issues often result in low accuracy, unreliable recognition rates, and slow convergence. To address the challenges posed by non-frontal visual features, this study proposes a hybrid model that fusion of Convolutional Neural Networks (CNN) and Bidirectional Long Short-Term Memory (BiLSTM), augmented with a point-multiplication attention model and Linear Discriminant Analysis (LDA). Data preparation incorporates median filtering and global contrast normalization to enhance image quality. The model then utilizes DenseNet and Softmax for image reconstruction, improving feature maps and providing essential data for classification tasks on the FER2013 and CK+ datasets. We benchmark our proposed model against conventional models such as Convolutional Neural Networks- Long Short-Term Memory (CNN-LSTM), Depthwise Separable Convolutional Neural Networks- Long Short-Term Memory (DSCNN-LSTM), Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-BiLSTM), and Attention Convolutional Neural Network - Long Short-Term Memory (ACNN-LSTM), evaluating performance metrics including F1 score, accuracy, precision, and recall. The results demonstrate that our proposed model outperforms existing approaches, highlighting the effectiveness of incorporating attention paradigms, hybrid deep learning architectures, and advanced preprocessing methods for facial emotion detection. The non-parametric statistical test also analyzes it.

Receive Date

27/11/2024

Revise Date

13/01/2025

Accept Date

16/01/2025

Publication Date

2-20-2025

References

[1] Wang S, Shuai H, Zhu L, Liu Q. Expression complementary disentanglement network for facial expression recognition. Chin J Electron 2024;33(3):742e52. https://doi.org/10.23919/ cje.2022.00.351.

[2] Raman DR, Kumar V, Pillai BG, Rabadiya D, Patre S, Meenakshi R. Multi-modal facial expression recognition through a hierarchical cross-attention graph convolutional network. In: 2024 international conference on knowledge engineering and communication systems (ICKECS), Chikkaballapur, India; 2024. p. 1e5. https://doi.org/10.1109/ ICKECS61492.2024.10616566.

[3] https://www.animeoutline.com/12-anime-facial-expressionschart-drawing-tutorial/.

[4] Jeremiah SR, Ha J, Singh SK, Park JH. Articles privacy guard: collaborative edge-cloud computing architecture for attribute-preserving face anonymization in CCTV networks. Human-centric Comput Informat Sci 2024;14(43):1e16.

[5] Kumar, A., Singh, S. K., Ravikumar, R. N., Khanna, A., & Brahma, B. Fusion of DRL and CNN for effective face recognition. Inform Syst Design: AI and ML Appl, 129.

[6] Kurde A, Singh SK. Next-generation technologies for secure future communication-based social-media 3.0 and smart environment. IECE Trans Sens, Communicat Control 2024; 1(2):101e25.

[7] Singh SK, Kumar M, Khanna A, Virdee B. Blockchain and FL-based secure architecture for enhanced external intrusion detection in smart farming. IEEE Internet Things J 2024;12(3): 3297e304. https://doi.org/10.1109/JIOT.2024.3478820.

[8] C P S, Ml ASJ. Integrated facial expression recognition on occluded faces using feature fusion. In: 2024 3rd international conference on sentiment analysis and deep learning (ICSADL), Bhimdatta, Nepal; 2024. p. 451e6. https://doi.org/ 10.1109/ICSADL61749.2024.00079.

[9] Almulla MA. Facial expression recognition using deep convolution neural networks. In: 2024 IEEE annual congress on artificial intelligence of things (AIoT), Melbourne, Australia; 2024. p. 69e71. https://doi.org/10.1109/AIoT63253. 2024.00022.

[10] Abu Mangshor NN, Ishak NH, Zainurin MH, Rashid NAM, Mohd Johari NF, Sabri N. Implementation of facial expression recognition (FER) using convolutional neural network (CNN). In: 2024 IEEE 15th control and system graduate research colloquium (ICSGRC), SHAH ALAM, Malaysia; 2024. p. 92e6. https://doi.org/10.1109/ICSGRC62081.2024.10691228.

[11] Yang D, Jiang S, Wang W. Research on facial expression recognition algorithm based on improved StarGAN network. In: 2024 2nd International Conference on signal processing and intelligent computing (SPIC), Guangzhou, China; 2024. p. 852e5. https://doi.org/10.1109/SPIC62469. 2024.10691554.

[12] Wang R, Ji Y, Li J. ASD diagnosis using facial expression recognition and gaze estimation. In: 2024 3rd international joint conference on information and communication engineering (JCICE), Fuzhou, China; 2024. p. 192e6. https:// doi.org/10.1109/JCICE61382.2024.00047.

[13] Aggarwal S. Comparative analysis of hedge funds in financial markets using machine learning models. Int J Comput Appl 2017;163(3).

[14] Li M, Jiang F, Zhang S. Research on facial expression recognition in the case of occlusion. In: 2024 5th international conference on computer vision, image and deep learning (CVIDL), Zhuhai, China; 2024. p. 328e33. https://doi.org/ 10.1109/CVIDL62147.2024.10603506.

[15] Chumachenko K, Iosifidis A, Gabbouj M. MMA-DFER: MultiModal adaptation of unimodal models for dynamic facial expression recognition in-the-wild. In: 2024 IEEE/CVF conference on computer vision and pattern recognition workshops (CVPRW), Seattle, WA, USA; 2024. p. 4673e82. https://doi.org/10.1109/CVPRW63382.2024.00470.

[16] Sambare Manas. FER-2013 Learn facial expressions from an image. November 2024. access on, https://www.kaggle.com/ datasets/msambare/fer2013.

[17] Yang G, Wang G, Li Y, Yu W. ACBL: attentive CNN-BiLSTM model for trajectory prediction. In: 2024 43rd Chinese control conference (CCC), Kunming, China; 2024. p. 8243e8. https:// doi.org/10.23919/CCC63176.2024.10662836.

[18] Meghana ML, Lakshmi KSV, Harshit NC, Vani KS. Cardiovascular disease detection in ECG images using CNNBiLSTM model. In: 2024 IEEE international conference on information technology, electronics and intelligent communication systems (ICITEICS), Bangalore, India; 2024. p. 1e6. https://doi.org/10.1109/ICITEICS61368.2024.1062 5411.

[19] Yao X, Lv Z, Cao L, Jiang F, Shan T, Wang J. Parallel CNNBiLSTM fault diagnosis method based on multi-domain transformation. In: 2024 IEEE international conference on sensing, diagnostics, drognostics, and control (SDPC), Shijiazhuang, China; 2024. p. 42e6. https://doi.org/10.1109/ SDPC62810.2024.10707762.

[20] Xu J, Zeng P. Short-term load forecasting by BiLSTM model based on multidimensional time-domain feature. In: 2024 4th international conference on neural networks, information and communication engineering (NNICE), Guangzhou, China; 2024. p. 1526e30. https://doi.org/10.1109/NNICE 61279.2024.10498827.

[21] Duan A, Raga RC. BiLSTM model with Attention mechanism for multi-label news text classification. In: 2024 4th international conference on neural networks, information and communication engineering (NNICE), Guangzhou, China; 2024. p. 566e9. https://doi.org/10.1109/NNICE61279.2024.10498894.

[22] Dino Hivi I, Abdulrazzaq Maiwan B. A comparison of four classification algorithms for facial expression recognition. Polytec J 2020;10(1):13. https://doi.org/10.25156/ptj.v10n1y2020.pp74-80.

[23] Hamad Zana O. Review of feature selection methods using optimization algorithm (Review paper for optimization algorithm). Polytec J 2023;12(2):24. https://doi.org/10.25156/ ptj.v12n2y2022.pp203-214.

Download

Share

COinS