Deep facial emotion recognition in video using eigenframes

Hajarolasvadi, Noushin; Demirel, Hasan

doi:10.1049/iet-ipr.2019.1566

Cited by 17 publications

(4 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lekdioui et al, [20] proposed an approach to identifying facial expressions based on the texture and shape descriptors of the face. CNN can yield good results when trained to analyse a face and recognize the features that influence its predictions [21]. This factor is crucial in obtaining positive outcomes from CNN.…”

Section: An Analysis Of Prior Research In the Relevant Fieldmentioning

confidence: 99%

The Implementation of Transfer Learning by Convolution Neural Network (CNN) for Recognizing Facial Emotions

Anbananthan Pillai Munanday,

Norazlianie Sazali,

Arjun Asogan

et al. 2023

ARASET

View full text Add to dashboard Cite

The primary objective of this study is to develop a real-time system that can predict the emotional states of an individual who commonly undergoes various experiences. The primary methodology suggested in this research for detecting facial expressions involves the integration of transfer learning techniques that incorporate convolutional neural networks (CNNs), along with a parameterization approach that minimizes the number of parameters. The FER-2013, JAFFE, and CK+ datasets were jointly used to train the CNN architecture for real-time detection, which broadened the range of emotional expressions that may be recognized. The proposed model has the capability to identify various emotions, including but not limited to happiness, fear, surprise, anger, contempt, sadness, and neutrality. Several methods were employed to assess the efficacy of the model's performance in this study. The experimental results indicate that the proposed approach surpasses previous studies in terms of both speed and accuracy.

show abstract

Section: An Analysis Of Prior Research In the Relevant Fieldmentioning

confidence: 99%

The Implementation of Transfer Learning by Convolution Neural Network (CNN) for Recognizing Facial Emotions

Anbananthan Pillai Munanday,

Norazlianie Sazali,

Arjun Asogan

et al. 2023

ARASET

View full text Add to dashboard Cite

show abstract

“…On the other hand, some researchers have had difficulty recognizing emotions accurately due to personal emotions' ambiguity and multifaceted nature [5,6]. Researchers have attempted to recognize and detect emotions using speech recognition [7,8], violence analysis from images [9][10][11], and Natural Language Processing (NLP), which are generally applied to large datasets created using social media. In order to understand the emotion of interactions, applications such as Instagram, Facebook, Twitter, and Youtube were used [12][13][14][15].…”

Section: Introductionmentioning

confidence: 99%

BP19: Tek Boyutlu İkili Modele Dayalı Doğru Bir Sesli Şiddet Tespit Modeli

YILDIZ

Keleş

YILDIRIM

et al. 2023

Turkish Journal of Science and Technology

View full text Add to dashboard Cite

Background: Audio violence detection (AVD) is a hot-topic research area for sound forensics but there are limited AVD researches in the literature. Our primary objective is to contribute to sound forensics. Therefore, we collected a new audio dataset and proposed a binary pattern-based classification algorithm. Materials and method: In the first stage, a new AVD dataset was collected. This dataset contains 301 sounds with two classes and these classes are violence and nonviolence. We have used this dataset as a test-bed. A feature engineering model has been presented in this research. One-dimensional binary pattern (BP) has been considered to extract features. Moreover, we have applied tunable q-factor wavelet transform (TQWT) to generate features at both frequency and space domains. In the feature selection phase, we have applied to iterative neighborhood component analysis (INCA) and the selected features have been classified by deploying the optimized support vector machine (SVM) classifier. Results: Our model achieved 97.01% classification accuracy on the used dataset with 10-fold cross-validation. Conclusions: The calculated results clearly demonstrated that feature engineering is the success solution for violence detection using audios.

show abstract

“…Given the ambiguity and versatility of personal emotions, their accurate recognition becomes an extremely challenging task which has been investigated by several researchers [3,4]. Researchers have attempted to detect and recognise emotions using speech recognition [3,5,6], image analysis [7,8,9] and Natural Language Processing (NLP) which is often applied to the expansive datasets formed through the use of social media services such as Facebook, Twitter, and YouTube to understand the sentiment of the interaction [10,11,12,13]. The ultimate objective of NLP is to interpret, decipher, and make sense of the human languages in a manner that is valuable.…”

Section: Introductionmentioning

confidence: 99%

DeepSafety:Multi-level Audio-Text Feature Extraction and Fusion Approach for Violence Detection in Conversations

Anwar¹,

Kanjo²,

Anderez³

2022

Preprint

View full text Add to dashboard Cite

Natural Language Processing has recently made understanding human interaction easier, leading to improved sentimental analysis and behaviour prediction. However, the choice of words and vocal cues in conversations presents an underexplored rich source of natural language data for personal safety and crime prevention. When accompanied by audio analysis, it makes it possible to understand the context of a conversation, including the level of tension or rift between people. Building on existing work, we introduce a new information fusion approach that extracts and fuses multi-level features including verbal, vocal, and text as heterogeneous sources of information to detect the extent of violent behaviours in conversations. Our multilevel multimodel fusion framework integrates four types of information from raw audio signals including embeddings generated from both BERT and Bi-long short-term memory (LSTM) models along with the output of 2D CNN applied to Melfrequency Cepstrum (MFCC) as well as the output of audio Time-Domain dense layer. The embeddings are then passed to three-layer FC networks, which serve as a concatenated step. Our experimental setup revealed that the combination of the multi-level features from different modalities achieves better performance than using a single one with F1 Score=0.85. We expect that the findings derived from our method provides new approaches for violence detection in conversations.

show abstract

Deep facial emotion recognition in video using eigenframes

Cited by 17 publications

References 38 publications

The Implementation of Transfer Learning by Convolution Neural Network (CNN) for Recognizing Facial Emotions

The Implementation of Transfer Learning by Convolution Neural Network (CNN) for Recognizing Facial Emotions

BP19: Tek Boyutlu İkili Modele Dayalı Doğru Bir Sesli Şiddet Tespit Modeli

DeepSafety:Multi-level Audio-Text Feature Extraction and Fusion Approach for Violence Detection in Conversations

Contact Info

Product

Resources

About