| Yazarlar (1) |
Dr. Öğr. Üyesi Senem TANBERK
Huawei R&D İstanbul, Türkiye |
| Özet |
| Keyframe detection is the study of selecting scenes that can best summarize long videos. Extracting a video summary is an important task to facilitate quick browsing and summarizing content, or for use in pre-processing in some image processing methods. The resulting photos are used for automated tasks in different industries (eg summarizing security footage, identifying different scenes used in music clips). In addition, processing high-volume videos in advanced machine learning methods creates resource costs. Keyframes obtained; It can be used as an input feature to the methods and models to be used.In this study; A deep learning-based approach is proposed for keyframe detection using a deep auto-encoder model with an attention layer. In the proposed method, first the features are extracted from the video frames using the encoder part of the autoencoder, and segmentation is applied using the k-means clustering algorithm to group similar frames together with these features. Then, the squares closest to the center of the clusters are determined and keyframes are selected from each cluster. The method was evaluated on the TVSUM video dataset and achieved a classification accuracy of 0.77, indicating a higher success rate than many existing methods. The proposed method offers a promising solution for keyframe extraction in video analysis and can be applied to a variety of applications such as video summarization and video retrieval. |
| Anahtar Kelimeler |
| Makale Türü | Özgün Makale |
| Makale Alt Türü | Uluslararası alan indekslerindeki dergilerde yayınlanan tam makale |
| Dergi Adı | x |
| Dergi Tarandığı Indeksler | |
| Makale Dili | İngilizce |
