The Capttitude is a model that is capable of generating affective image captions with an emotional component and thereby going beyond the factual image descriptions. It provides access to two methods. Firstly, combination of a Convolutional Neural Network (CNN) with a Long Short Term Memory (LSTM) Network [Show and Tell]. The output sentences are further refined with the addition of ANPs from DeepSentiBank. Secondly, a graph based Concept and Syntax Transition (CAST) Network. The directed graph is generated from the training set captions of YFCC100M dataset, connecting the “Concepts” (nouns, adjectives, verbs) with one another. To expand the vocabulary of the captioning model Word2Vec similarity is used to connect conceptually similar nodes in the graph. A run through the graph from “Start” to “End” containing the activated nodes represents a generated caption of the image. Hence, a user is provided with a pair of emotionally rich captions from these two methods.

link to demo contact

Semiotics is the study of sign processes and meaningful communication.

Video Semiotics extracts high level semantics from video. It not only detects objects but it also associates a sentiment through a Deep Neural Network trained on Adjective Noun Pairs (ANPs). Contact

Video Samples:

TubeTagger is a concept-based video retrieval system that learns to detect visual concepts (like “soccer”, “desert”, or “interview”) automatically from YouTube. The demo allows you to search a video collection using keywords. The videos have been indexed completely automatically by TubeTagger - no manual tagging was done!

link to demo MOONVID project contact

Smart Video Buddy is an intelligent assistant that understands videos and links them with other content. Semantic concepts (like “soccer game”) are detected in real-time, and this information is used to “make videos smarter”, i.e. to enrich them with adapted news, advertisements, or interesting links. Smart Video Buddy was presented at the CeBIT 2010.

link to demo demo flyer MOONVID project contact

The InViRe system realizes content-based retrieval in a dataset of TV content. If you click on a scene, InViRe uses video fingerprints based on a video's color, texture, and motion to retrieve similar scenes.

link to demo contact

This demo illustrates the automatic recognition of pornographic material using computer vision techniques (with a click, you can re-rank material in a database such that potentially offensive content is shown). We developed the system as part of the FIVES project, in which we develop a toolbox supporting police investigators with the detection of illegal child pornographic material.

link to demo FIVES project contact

The Navidgator online demo allows the structural browsing of a image/video database in respect to the visual similarity of the documents.

link to TV demo & image demo contact

Document security is gaining more and more importance in every day life. The widespread availability of scanners and printers allows even untrained people to easily forge and alter documents. The GuardDoc demo system shows the application of automatic layout comparison technology for document security applications.

link to demo video contact

This demo shows the application of automatic line orientation measurement for document security applications.

link to demo contact

To limit the fraudulent use of color laser printers and color copiers, the so called Counterfeit Protection System (CPS) codes (“yellow dots”, “tracking dots”) have been introduced. Despite the fact that most information about these patterns is classified, we were able to develop image based methods to identify print outs from the same printer of the same printer class.

link to demo overview contact

OCRopus™ is a state-of-the-art document analysis and OCR system, featuring pluggable layout analysis, pluggable character recognition, statistical natural language modeling, and multi-lingual capabilities. The system is being developed with the generous support from Google and other organizations; the primary developers are at the IUPR Research Group.

link to demo OCRopus project homepage contact

Layout analysis remains as a significant performance limiting step in OCR and document analysis systems. To make progress in this area, we have developed a set of benchmarking tools and tasks and applied them to the evaluation of different document layout analysis methods.

link to demo contact

This is an online demo of our IP geo location software. The software generates a IP-to-country lookup table out of the Regional Internet Registrars (RIR) databases, checks plausibility and resolves conflicts. A recent database can be found on the project's website.

link to demo
project website
contact: Markus Goldstein

If you are under a Distributed Denial of Service (DDoS) attack, you need to filter out the attacker's traffic in order to get your server online again. If attacks are highly distributed, this is a very difficult task. This demo generates out positive access rules for your firewall based on the log-files of your server (e.g. Apache). If you enable the firewall ruleset, the attack will be blocked minimizing blocking your valid customers.

link to demo
network security website
contact: Markus Goldstein