- د.محمد الجراح و د. عبدالكريم التميمي
- بتمويل من جامعة اليرموك
- 7400 دينار اردني
The project aims to help solve one of the main challenges in multimedia systems and Big Data fields. With the recent explosion of video traffic and video contents, it becomes harder to index and retrieve videos based on their contents. It is even harder to find similar videos to a certain video of interest.
Our proposed system aims to facilitate the process of content abstraction using KAZE feature descriptor, and index similar video features in a relatively close distance using locality sensitive hashing (LSH). This approach allows us to query videos based on their extracted features and retrieve similar videos with high accuracy and relatively low resources requirements.
Our proposed system will base its decision on the features extracted from more than 1 million raw images.