A context inclusive approach to face detection in video the effect of temporal context

No Thumbnail Available
Issue Date
Journal Title
Journal ISSN
Volume Title
In this bachelor thesis we have investigated the e ects of temporal context for face detec- tion in video. This idea has already been implemented in the eld of action recognition. Our contribution is exploring whether this idea is also relevant in the eld of face de- tection. This is realised by the addition of extra input images that vary in the time domain. The method that is used in ates lters in the rst layer of the neural network, enabling access to temporal features. We have researched whether this method will im- prove detection performance. However, the ndings in the scope of this thesis provide no evidence of a signi cant increase in detection performance when temporal context is added. Nevertheless, this research provides ndings indicative of potential gains in performance that can be expanded by future work, together with ideas that can bene t future research in this eld.
Faculteit der Sociale Wetenschappen