Zamansal video bölümleme

With the advancement in technology video content generation increased rapidly. This advancement of technology also increased the consumption of video information. The increase in both generation and consumption of video content has created the need of segmenting, summarizing and indexing video with high efficiency. Video segmentation is the first step to summarize and index videos. Video segmentation aims to segment a video into meaningful, consistent shots. After segmenting video into shots with consistent content then we can apply semantic segmentation techniques to further analysis of a video. In this thesis Temporal Video Segmentation is examined in both compressed and uncompressed domain and presented a new method using artificial neural networks that has improved performance over methods presented in the related work.