- Dataset 1 (or inspect individual files)
- Dataset 2 (or inspect individual files)
- Dataset 3 (or inspect individual files)
- Dataset 4 (or inspect individual files)
- Dataset 5 (or inspect individual files)
- Dataset 6 (or inspect individual files)
High Impact Visual Communication
Images can serve as a powerful communications vehicle, conveying a wealth of information as well as emotional impact. The color, composition, content, lighting and sharpness of an image all contribute to a viewer’s response to that image, and relative placement, scaling and orientation of a group of images in a collage adds an additional layer of richness and meaning to a page. These characteristics are used extensively by professionals on web sites, magazine covers and printed advertisements to draw attention, communicate a message and leave a lasting emotional impression.
Because images hold such power, people like to use their photos to tell their own stories. However, their end result often falls short since many people lack the skill and intuitive understanding needed to create a coherent visual story from their photos. In addition, a picture is worth a thousand words. How do we create a high impact picture that can convey information across cultural boundaries and find a thousand words that best describe such a picture? This grand challenge is to find a solution which can create a collage and generate a textual description that tells the story of the set of photos.
The system starts with a digital photo collection, such as photos taken during a vacation. It then analyzes the collection automatically using information from multiple sources such as image analysis, internet data sources, and EXIF tags. The result of the analysis is used to create the most appealing collage picture that best represents the original collection. In addition, it is also used to generate a description of the collage picture.
Data Set
We provide six datasets, each with 20 photos. The participants will develop a system that can use the photos in each dataset as input. For each of the dataset, the participants will generate a collage. One can choose to use any number of photos within the dataset to make the collage. The dimension of the collage will be 1600×1200 pixels. In addition, the participants will also use the system to generate a description (prose or paragraph in English) about the collage with 100 words or less. The system can access internet data sources while generating the results.
Evaluation
The evaluation of the submissions will be based on the following criteria:
- Relevance of the collage to the photos in the corresponding dataset
- The aesthetics of the collage
- Relevance of the description to the photos in the corresponding dataset
- The clarity and coherence of the description
- Originality of the algorithms for generating the collages and descriptions
- Speed performance of the system
Each team is encouraged to submit a collage picture (1600×1200 pixels) and a textual description (100 words or less) for each data set, as well as a written paper. Partial results will be accepted.
Feel free to correspond with the challenge authors via the comments form below.
For private correspondence, consult the here for contact details.