These graphs are for the IEEE/CVF Conference on Computer Vision and Pattern Recognition. The top graph is a visualization on the main conference papers for the conference distributed based on their similarity to each other. The closer the papers are, the more similar the abstracts. This graph can be used to search for papers and to find papers that are similar to each other. Once you find an interesting paper by searching, you can hover your mouse over nearby papers to see them. You can also click and drag a box over the graph to see a keyword summary and list of the papers on the right. The second graph shows connections between the keywords and subjects and the nodes can be picked up and moved to help analyze the graph. For more information, please visit our blog post.

Each dot represents a paper with the color representing the subject area (legend below this graph). The papers are arranged by a measure of similarity.

If you hover over a dot, you see the related paper.

If you click on a dot, you go to the related paper page.

You can search for papers by author, keyword, or title

Drag a rectangle to summarize an area of the plot.

Cutoff: 4

Each dot represents either a keyword (black) or a primary subject area (colorful). The subject area colors match those in the plot above.

You can hover over a dot to show the keyword or subject area associated with it.

You can click and drag a dot to move it around.

Each edge represents the keyword and subject area it connects being referenced by some number of papers--the edge thickness is proportional to this number.

Move the slider to hide keyword/area pairings with fewer than N referencing papers.

Graphs created by Dr. Christopher Funk (Kitware) and Dr. Roni Choudhury (Kitware). Special thanks to Dr. Hendrik Strobelt (IBM) for the original similarity graph created for Mini-Conf, Professor Sasha Rush (Cornell) for advice on the embeddings, and Dr. Anthony Hoogs (Kitware) and Dr. Jeffrey Baumes (Kitware) for support and advice. If you want to know more about the graphs, please visit our blog post.