Robust Approach for Emotion Classification Using Gait

Srivastava, Sahima; Rastogi, Vrinda; Prakash, Chandra; Sethi, Dimple

doi:10.1007/978-981-15-5113-0_74

Cited by 1 publication

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We evaluated our proposed model, referred to as AE-xyz against a CNN model, referred to as CNN-xyz, on the raw joint time-series data, and three shallow baseline models that have been used in affect recognition based on human motion data [19]. These are a K Nearest Neighbours (KNN) classifier and a Support Vector Machine classifier trained on the manually engineered angular and velocity features, referred to as KNNman and SVM-man, as well as an SVM-xyz model that was trained on the raw data.…”

Section: Evaluation Methodologymentioning

confidence: 99%

“…These are a K Nearest Neighbours (KNN) classifier and a Support Vector Machine classifier trained on the manually engineered angular and velocity features, referred to as KNNman and SVM-man, as well as an SVM-xyz model that was trained on the raw data. As a baseline the set of raw features extracted were the mean, std, min and max of the key angles such as knees or elbows, as well as velocities of key joints in the arms and legs, all of which have been shown to be useful features for affect recognition [19].…”

Section: Evaluation Methodologymentioning

confidence: 99%

“…On the other hand, there are also studies that use pose-based gait data to extract angular or velocity features. Srivastava et al [19] extracted angular features such as head tilt, shoulder, right elbow and right knee, and reported high accuracy with shallow machine learning models.…”

Section: E Manual Feature Extractionmentioning

confidence: 99%

See 2 more Smart Citations

Towards Explainable, Privacy-Preserved Human-Motion Affect Recognition

Malek-Podjaski

Deligianni

2021

2021 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

77Human motion characteristics are used to monitor the progression of neurological diseases and mood disorders. Since perceptions of emotions are also interleaved with body posture and movements, emotion recognition from human gait can be used to quantitatively monitor mood changes. Many existing solutions often use shallow machine learning models with raw positional data or manually extracted features to achieve this. However, gait is composed of many highly expressive characteristics that can be used to identify human subjects, and most solutions fail to address this, disregarding the subject's privacy. This work introduces a novel deep neural network architecture to disentangle human emotions and biometrics. In particular, we propose a cross-subject transfer learning technique for training a multi-encoder autoencoder deep neural network to learn disentangled latent representations of human motion features. By disentangling subject biometrics from the gait data, we show that the subject's privacy is preserved while the affect recognition performance outperforms traditional methods. Furthermore, we exploit Guided Grad-CAM to provide global explanations of the model's decision across gait cycles. We evaluate the effectiveness of our method to existing methods at recognizing emotions using both 3D temporal joint signals and manually extracted features. We also show that this data can easily be exploited to expose a subject's identity. Our method shows up to 7% improvement and highlights the joints with the most significant influence across the average gait cycle.

show abstract

Section: Evaluation Methodologymentioning

confidence: 99%

Section: Evaluation Methodologymentioning

confidence: 99%