BACHELOR Assignment
[M][B] Computer vision for image and video analysis
Type: Bachelor CS/EE/HMI/etc
Period: TBD
Student: (Unassigned)
If you are interested please contact :
Description:
Several projects available on this topic:
Computer vision for human related crime recognition
UCF-Crime is the largest dataset available for automatic visual anomaly analysis and consists of real-world crime scenes of various categories. Recently, HR-Crime was introduced as a subset of the UCF-Crime dataset suitable for human-related anomaly detection tasks. Previous work in this field relied on descriptors such as skeleton trajectories, video depth, audio signals and radar for the recognition of different human activities.
In this project, you are expected to build a framework for the classification of videos based on visual descriptors.
Related works:
- Boekhoudt, Kayleigh, et al. "HR-Crime: Human-Related Anomaly Detection in Surveillance Videos." International Conference on Computer Analysis of Images and Patterns. Springer, Cham, 2021.
Computer vision for behaviour analysis
Human activity analysis has a wide range of real-world applications - health monitoring, human-computer interaction, routine analysis, to name some. Video-based activity recognition usually entails the detection and classification of spatial-temporal behavioural patterns. Previous works in this field relied on descriptors such as skeleton trajectories, video depth, audio signals, and radar for the recognition of different human activities.
In this project, you are expected to build a framework for the classification of videos based on visual descriptors.
Related works:
- Glavan, A., & Talavera, E. (2022). InstaIndoor and multi-modal deep learning for indoor scene recognition. Neural Computing and Applications, 34(9), 6861-6877.
- Matei, A., Glavan, A., Radeva, P., & Talavera, E. (2021). Towards Eating Habits Discovery in Egocentric Photo-Streams. IEEE Access, 9, 17495-17506.
- Costa, W., Macêdo, D., Zanchettin, C., Figueiredo, L. S., & Teichrieb, V. (2021). Multi-Cue Adaptive Emotion Recognition Network. arXiv preprint arXiv:2111.02273.
Other topics available are related to medical image analysis, event-based cameras for video analysis, and sentiment analysis from visuals and text, among others. Projects that work with deep learning architectures such as CNNs and Transformers for visual data analysis are also interesting.