Our demo paper Finding the Chameleon in Your Video CollectionÂ presented at the ACM Multimedia Systems 2016 Conference (MMSYS 2016) is now available for download. It describes an interactive video retrieval tool that is optimized for search by different typesÂ (sketch, example segment, and semantic concepts) and makes uses CNNs (convolutional neural networks) as well as optical flow estimation andÂ temporal features signatures to analyze the content in the video. The interface is designed according to research results from several years on content-based search in video.
The Next Generation Video Browsing (NGVB) project, an FWF funded translational research project (TRP 273-N15), which ran for three years, has ended in January this year. In the NGVB project we investigated how to design better tools for content-based search in videos, in particular when used on tablet devices, and how these search tools could benefit from the powerful features provided by tablets (e.g., multi-touch interaction and high computing power). We focused on improving video content navigation and visualization of the video structure through 3D models and color-sorted arrangement of images. To this end, we performed several initial user studies in order to find out how users search with common video player software and used the results for the design of novel search features. In order to evaluate these novel search features we developed software tools with new interaction models and search features and performed comparative user studies with them. The results of these studies show that the newly proposed features and interfaces provide significantly better search performance and are easier to use for the vast majority of tested users.
The research work of the NGVB project brought forward several interesting results and I would like to thank all involved people (in particular Dr. Marco A. Hudelist and Dr. Claudiu Cobarzan). The figure below shows an overview of research prototypes (i.e., interfaces) for video and image browsing that evolved from the project. More details can be found on our project website NGVB.net (including related research papers).
We are very happy to announce that our paper on
“Domain-Specific Video Compression for Long-term Archiving of Endoscopic Surgery Videos“,
authored by Bernd MĂĽnzer, Klaus Schoeffmann, and Laszlo BĂ¶szĂ¶rmenyi,
has been accepted to the 29th IEEE International Symposium on Computer-Based Medical Systems (CBMS 2016), to be held in June 2016 in Dublin/Belfast.
A preprint of the paper is available here.
Congratulations again to Kai Uwe Barthel, Nico Hezel, and Radek Mackowiak (and Florian Barthel), from HTW Berlin, Germany, for winning the Video Browser Showdown 2016 competition with their interactive video search system described in Navigating a graph of scenes for exploring large video collections!
Multimedia information indexing and retrieval nowadays is more and more penetrating an important domain for society: healthcare. Feature-based classification approaches which are being developed for medical image classification for computer-aided diagnosis borrow the approaches from classical CBIR in feature engineering and cascaded classification. Deep learning classifiers which are being extensively studied and applied for concept recognition in multimedia data, image and video understanding are being applied for prediction of patients categories on the basis of physiological parameters such as gaze fixations. Information fusion approaches which are necessary for understanding and content â€“ based indexing of highly dimensional multimedia data are applied for fusion of different modalities in medical image recognition. Video analysis and summarization approaches are being developed for automatic visual reporting in surgery. Similarly, video content analysis and retrieval in archived video data collected from surgeries becomes more and more important and provides the basis for later usage of these valuable data for scenarios such as case comparisons/similarity search, teaching of new operation techniques, as well as quality control/error inspection. Finally, the multimedia nowadays is more and more multimodal â€“ not only image, video, textual and sound modalities supply the information, but also and specifically in medical and healthcare applications, a large variety of different sensors either measuring the context or physiological parameters are deployed. Future multimedia becomes multimodal and this happens in the healthcare domain in priority.
We are now looking for papers (6 pages, IEEE style) for this exciting special session (deadline: February 1, 2016).
More information can be found here.
The intimate presence of mobile devices in our daily life, such as smartphones and various wearable gadgets like smart watches, has dramatically changed the way we connect with the world around us. Users rely on mobile devices to maintain an always-on relation to information, personal and social networks, etc. With a growing number of powerful embedded mobile sensors like camera, microphone, GPS, gyroscope, accelerometer, digital compass, and proximity sensor, there is a variety of data available and hence enables new sensing applications across diverse research domains comprising mobile media analysis, mobile information retrieval, mobile computer vision, mobile social networks, mobile human-computer interaction, mobile entertainment, mobile gaming, mobile healthcare, mobile learning, and mobile advertising.
Regardless of the application fields, many issues and challenges brought by the emerging technologies for mobile multimedia still lie ahead and many research questions remain to be answered. For example, seamless user experience has been identified as one key factor in designing mobile multimedia applications for multiple form factors. Yet, its provision is challenging and requires effective integration of rich mobile sensors and multidisciplinary research, such as multimedia content adaptation and user behavior analysis. Also, the effective and efficient use of multimedia data on mobile devices, including convenient interaction and appropriate visualization, is still not fully solved. Small screen sizes, integrated cameras and microphones, but also the new gesture and haptic sensors, pose new challenges but at the same time provide new opportunities and innovative ways for multimedia interaction. In addition, for power saving purpose, application-driven energy management is an important technical consideration in mobile multimedia computing (e.g., how to offload computation-intense tasks to the cloud servers to save energy and extend battery lifetimes for mobile users?).
The workshop on Mobile Multimedia Computing (MMC 2016) aims to bring together researchers and professionals from worldwide academia and industry for showcasing, discussing, and reviewing the whole spectrum of technological opportunities, challenges, solutions, and emerging applications in mobile multimedia.
See the call for papers!
HealthWear’16 will bring together researchers, developers, and industry professionals from both Healthcare and Quantified Self communities to discuss key issues, opportunities and obstacles for personal health data research. These include challenges of capturing, summarizing, presenting and retrieving relevant information from heterogeneous sources to support a new vision of pervasive personal healthcare.
The conference now calls for tutorials focusing on practical topics related to wearables and healthcare. Current topics of interest as found in the HealthWear’16 CFP are:
- Personal Health Informatics
- Quantified Self for Healthcare
- Activity Monitors and Devices
- Healthcare Knowledge Representation & Reasoning
- Health Data acquisition, analysis and mining
- Healthcare Information Systems
- Validity, reliability, usability, and effectiveness of Self-Tracking devices
- Experiment Design
- Social and Psychological investigation into Self-Tracking practices
- Health Monitoring in clinical and lifestyle environments
- Sensors and actuators for Wellness, Fitness and Rehabilitation
- Innovative Algorithms for assessment of long-term physiological and behavioural data
- Models for interpreting medical sensor data
- Lifelogging, lifecaching, lifestreaming
- Biometric data
- Medical Self-diagnostics
HealthWear’16 solicits proposals for half-day tutorials that present fundamentals about a frontier topic as well as recent advances in fields mentioned above. It should be interesting for both students and senior researchers. Of particular interest are topics focusing on the practical challenges of the Quantified Seld (QS), such as visualization of QS data, search in this special kind of multi modal data, etc.
More information can be found here.
The slides of the tutorial on Interactive Video Search, presented yesterday at the ACM International Conference on Multimedia 2015 (ACM MM’15) in Brisbane, Australia, can be found here:
Our surveyÂ paper on video interaction tools is now available for download atÂ the ACM portal website. It reviews a large numberÂ of tools andÂ approaches for video content interaction, proposed by colleagues over the last fewÂ years.
Klaus Schoeffmann, Marco A. Hudelist, Jochen Huber, â€śVideo Interaction Tools: A Survey of Recent Workâ€ś.Â ACM Computing Surveys, Vol. 48, No. 1, ACM, September 2015, pp. 14:1-14:34,
We are happy to announce that our joint paper on “Endoscopic Video Retrieval: A Signature-based Approach for Linking Endoscopic Images with Video Segments” (authored together with Christian Beecks, Mathias Lux, Merih Seran Uysal, and Thomas Seidl) has been accepted to the IEEE International Symposium on Multimedia 2015 (ISM 2015), to be held in Miami, Florida, USA, in December 2015. Attached you can find a preprint of the paper.