An Affective Video Database using Multimedia Content Analysis rated on Indian samples

Mishra, Sudhakar; Asif, Mohammad; Srinivasan, Narayanan; Tiwary, Uma Shanker

doi:10.31234/osf.io/f5hj9

Cited by 3 publications

(5 citation statements)

References 26 publications

(34 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They used a combination of CNN and LSTM to extract spatiotemporal features using an urban surveillance video dataset. Furthermore, [21], [22] presented a deep learningbased approach to detect violent events in movie datasets.…”

Section: B Deep Learning-based Vd Approachesmentioning

confidence: 99%

VD-Net: An Edge Vision-Based Surveillance System for Violence Detection

Khan,

Saddik,

Gueaieb

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The automation of surveillance systems, driven by the rapid development of computer vision technology, has significantly enhanced the analysis of surveillance videos, particularly in recognition of human activity, including behavior analysis and violence detection, thereby bolstering public and industrial security. Despite these advancements, detecting and analyzing violent actions remains challenging, especially for real-time surveillance systems with limited computing power. We propose an artificial intelligencebased framework called VD-Net (Violence Detection Network), enabled by Intelligent Internet-of-Things (IIoT) to detect violent behavior in public and private spaces. The model utilizes lightweight special task temporal convolutional network (ST-TCN) blocks and several bottleneck layers to focus on salient features in the input sequence. The learned features passed from the classifier to discriminate between violent and nonviolent actions. Additionally, our system is supposed to trigger an alert if violence is detected, which is then communicated to relevant departments. We tested the effectiveness of the proposed system by conducting experiments on surveillance and non-surveillance datasets and ensured a 1-4 % improvement in State-of-The-Art (SoTA) accuracy.

show abstract

Section: B Deep Learning-based Vd Approachesmentioning

confidence: 99%

VD-Net: An Edge Vision-Based Surveillance System for Violence Detection

Khan,

Saddik,

Gueaieb

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The procedure for stimuli collection is explained elsewhere (Mishra et al, 2022). Briefly, we crawled YouTube to collect candidate emotional stimuli.…”

Section: Stimuli Collectionmentioning

confidence: 99%

“…After obtaining the scale ratings from 407 participants in the first stage of the validation study (from (Mishra et al, 2022)), we performed two operations. First, the emotion labels were tentatively assigned to stimuli to aid the stimuli selection for the second stage.…”

Section: Stimuli Selectionmentioning

confidence: 99%

“…We started our study with 222 multimedia video stimuli, which were obtained using multimedia content analysis described in (Mishra et al, 2022). Given the increasing amount of multimedia content on the internet, such a screening tool would be useful to make an initial selection of videos and manually handpicking a few videos can be potentially biased.…”

Section: Probability Of Eliciting An Emotional Clustermentioning

confidence: 99%

“…We reduced experimenter bias in the initial screening of emotional stimuli by relying on computer algorithms to search multimedia videos for affective terms, followed by automatic content analysis to calculate the valence and arousal from audio-visual features, which are subsequently utilized to select the candidate film clips (Mishra, Srinivasan, & Tiwary, 2022). Rather than restricting to a small number of emotional categories, we asked the participants about the emotion category they felt during viewing clips to reduce experimenter bias further.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Affective film dataset from India (AFDI): creation and validation with an Indian sample

Mishra,

Srinivasan,

Asif

et al. 2023

J Cult Cogn Sci

View full text Add to dashboard Cite

We describe the creation of an affective film dataset for researchers interested in studying a spectrum of emotional experiences. We followed a two stage process. In the first stage, two hundred twenty-two video clips with 60-seconds long duration were rated in the lab by 407 participants. Based on the selection criteria, 69 audio-visual clips were selected for the second stage. These selected affective clips are then presented to 271 participants. Participants rated these clips on rating scales and categorized them into emotion categories. We clustered these emotional ratings into six categories, calculated the probability of eliciting emotion categories for each stimulus and compared the probability results with datasets created with the Western population. Given that part of the dataset consisting of videos from the west have been rated earlier, we compared ratings by Indian and Western participants on that subset. Even with some differences in methodology and analyses, we did 1 AFDI 2 AFDI not observe significant differences in ratings of different scales. Interestingly, the ratings to Indian stimuli by Indian participants tended to be significantly higher in arousal, dominance and familiarity than the ratings to western stimuli by Indian participants. The dataset we have created comprises film clips based on Indian and western content, making it flexible for use in Indian or western cultures. We hope that this dataset emotion researchers with a more diverse set of emotional movie clips accompanied by emotional ratings and other quantitative measures.

show abstract

manuscript

Mishra¹,

Tiwary²,

Srinivasan³

et al. 2022

Preprint

View full text Add to dashboard Cite

We describe the creation of an affective film dataset for researchers interested in studying a spectrum of emotional experiences. We followed a two stage process. In the first stage, two hundred twenty-two video clips with 60-seconds long duration were rated in the lab by 407 participants. Based on the selection criteria, 69 audio-visual clips were selected for the second stage. These selected affective clips are then presented to 271 participants. Participants rated these clips on rating scales and categorized them into emotion categories. We clustered these emotional ratings into six categories, calculated the probability of eliciting emotion categories for each stimulus and compared the probability results with datasets created with the Western population. Given that part of the dataset consisting of videos from the west have been rated earlier, we compared ratings by Indian and Western participants on that subset. Even with some differences in methodology and analyses, we did not observe significant differences in ratings of different scales. Interestingly, the ratings to Indian stimuli by Indian participants tended to be significantly higher in arousal, dominance and familiarity than the ratings to western stimuli by Indian participants. The dataset we have created comprises film clips based on Indian and western content, making it flexible for use in Indian or western cultures. We hope that this dataset emotion researchers with a more diverse set of emotional movie clips accompanied by emotional ratings and other quantitative measures.

show abstract

An Affective Video Database using Multimedia Content Analysis rated on Indian samples

Cited by 3 publications

References 26 publications

VD-Net: An Edge Vision-Based Surveillance System for Violence Detection

VD-Net: An Edge Vision-Based Surveillance System for Violence Detection

Affective film dataset from India (AFDI): creation and validation with an Indian sample

manuscript

Contact Info

Product

Resources

About