The paper (titled “Video Matching Using DC-image and Local Features”) was presented by Saddam Bekhet (PhD Rsearcher) in the International Conference of Signal and Image Engineering (ICSIE’13), during the World Congress on Engineering 2013, in London UK.
Abstract:
This paper presents a suggested framework for video matching based on local features extracted from the DC-image of MPEG compressed videos, without decompression. The relevant arguments and supporting evidences are discussed for developing video similarity techniques that works directly on compressed videos, without decompression, and especially utilising small size images. Two experiments are carried to support the above. The first is comparing between the DC-image and I-frame, in terms of matching performance and the corresponding computation complexity. The second experiment compares between using local features and global features in video matching, especially in the compressed domain and with the small size images. The results confirmed that the use of DC-image, despite its highly reduced size, is promising as it produces at least similar (if not better) matching precision, compared to the full I-frame. Also, using SIFT, as a local feature, outperforms precision of most of the standard global features. On the other hand, its computation complexity is relatively higher, but it is still within the real-time margin. There are also various optimisations that can be done to improve this computation complexity.
New Conference paper accepted for publishing in “World Congress on Engineering 2013“.
The paper title is “Video Matching Using DC-image and Local Features ”
Abstract:
This paper presents a suggested framework for video matching based on local features extracted from the DC-image of MPEG compressed videos, without decompression. The relevant arguments and supporting evidences are discussed for developing video similarity techniques that works directly on compressed videos, without decompression, and especially utilising small size images. Two experiments are carried to support the above. The first is comparing between the DC-image and I-frame, in terms of matching performance and the corresponding computation complexity. The second experiment compares between using local features and global features in video matching, especially in the compressed domain and with the small size images. The results confirmed that the use of DC-image, despite its highly reduced size, is promising as it produces at least similar (if not better) matching precision, compared to the full I-frame. Also, using SIFT, as a local feature, outperforms precision of most of the standard global features. On the other hand, its computation complexity is relatively higher, but it is still within the real-time margin. There are also various optimisations that can be done to improve this computation complexity.
(Click Semantic Video Annotation-with Knowledge ” https://amrahmed.blogs.lincoln.ac.uk/files/2013/03/Semantic-Video-Annotation-with-Knowledge.pdf , to download the pdf)
INTRODUCTION
The volume of video data is growing exponentially. This data need to be annotated to facilitate search and retrieval, so that we can quickly find a video whenever needed.
Manual Annotation, especially for such volume, is time consuming and would be expensive. Hence, automated annotation systems are required.
AIM
Automated Semantic Annotation of wide-domain videos (i.e. no domain restrictions). This is an important step towards bridging the “Semantic Gap” in video understanding.
METHOD
1. Extracting “Video Signature” for each video.
2. Match signatures to find most similar videos, with annotations
3. Analyse and process obtained annotations, in consultation with Common-sense knowledge-bases
4. Produce the suggested annotation.
EVALUATION
• Two standard, and challenging Datasets were used. TRECVID BBC Rush and UCF.
• Black-box and White-box testing carried out.
•Measures include: Precision, Confusion Matrix.
CONCLUSION
•Developed an Automatic Semantic Video Annotation framework.
•Not restricted to a specific domain videos.
•Utilising Common-sense Knowledge enhances scene understanding and improve semantic annotation.
Amjad Altadmri has passed his PhD viva, subject to minor amendments, earlier today.
Thesis Title: “Semantic Video Annotation in Domain-Independent Videos Utilising Similarity and Commonsense Knowledgebases”
Thanks to the external, Dr John Wood from the University of Essex, the internal Dr Bashir Al-Diri and the viva chair, Dr Kun Guo.
Congratulations and Well done.
All colleagues are invited to join Amjad on celebrating his achievement, tomorrow (Thursday 28th Feb) at 12:00noon, in our meeting room MC3108, with some drinks and light refreshments available.
New Journal paper accepted for publishing in the Journal of “Multimedia Tools and Applications“.
The paper title is “A Framework for Automatic Semantic Video Annotation utilising Similarity and Commonsense Knowledgebases”
Abstract:
The rapidly increasing quantity of publicly available videos has driven research into developing automatic tools for indexing, rating, searching and retrieval. Textual semantic representations, such as tagging, labelling and annotation, are often important factors in the process of indexing any video, because of their user-friendly way of representing the semantics appropriate for search and retrieval. Ideally, this annotation should be inspired by the human cognitive way of perceiving and of describing videos. The difference between the low-level visual contents and the corresponding human perception is referred to as the ‘semantic gap’. Tackling this gap is even harder in the case of unconstrained videos, mainly due to the lack of any previous information about the analyzed video on the one hand, and the huge amount of generic knowledge required on the other.
This paper introduces a framework for the Automatic Semantic Annotation of unconstrained videos. The proposed framework utilizes two non-domain-specific layers: low-level visual similarity matching, and an annotation analysis that employs commonsense knowledgebases. Commonsense ontology is created by incorporating multiple-structured semantic relationships. Experiments and black-box tests are carried out on standard video databases for
action recognition and video information retrieval. White-box tests examine the performance of the individual intermediate layers of the framework, and the evaluation of the results and the statistical analysis show that integrating visual similarity matching with commonsense semantic relationships provides an effective approach to automated video annotation.
