Supplementary file: AverageExplorer: Interactive Exploration and Alignment of Visual Data Collections

2014

Citations in the corpus

• Generative Visual Manipulation on the Natural Image Manifold (2016, background): “The difference with previous sketch-to-image retrieval methods [41] or AverageExplorer [42], is that due to potentially contradicting user constraints, the result may look very different than any single image from the dataset or an average of such images, and more of a realistic hybrid image [43].”
• Attribute2Image: Conditional Image Generation from Visual Attributes (2015, methodology): “As suggested by Zhu et al. [44], average image of aligned K-nearest neighbors captures better visual similarity to ground-truth than K-nearest neighbors, which we added as a strong baseline (NN avg ).”
• GAN-Supervised Dense Visual Alignment (2021, methodology): “Joint visual alignment and clustering was proposed in AverageExplorer [95] but as a user-driven data interaction tool.”
• Sketch Me That Shoe (2016, methodology): “Other SBIR works like Sketch2Photo [4] and AverageExplorer [34], use sketch in addition to text or colour cues for image retrieval.”
• LiveSketch: Query Perturbations for Guided Sketch-Based Visual Search (2019, unknown): “Our work is aligned to ShadowDraw [22] in which ghosts (edge-maps derived from top search results) are averaged and overlaid onto the sketch canvas (similarly, [40] for photo search).”
• Dataset fingerprints: Exploring image collections through data mining (2015, methodology): “An interactive method that allows a user to explore and visualize a large image collection using simple average images is presented in [27].”