The Challenge
Given a selection of images, and choosing a particular reference one, find the 3 “closest” images to that reference image. How “close” an image is determined by the number of pixels that match precisely between the two images - both in value and location. % Closeness = (Number of Pixels Matching) / (Number of Pixels Total) * 100 The output should be a list of images, and their to the reference either as full percentage (97%) or decimal (0.97)
The Images
Images are purely random noise. They are 1024 pixels wide and 1024 pixels high. Each pixel is encoded as 3 bytes of data, representing the Red, Green and Blue Channels. For example: the first byte could be represented as [255, 255, 255] which would be a white pixel. Another pixel could be [0, 0, 0] which would be a black pixel.
Levels
As with all games, there are levels!
Bronze - 10 Images in dataset (30 megabytes) https://pixels-data.s3.eu-west-2.amazonaws.com/Bronze.zip
Silver - 100 Images in dataset (300 megabytes) https://pixels-data.s3.eu-west-2.amazonaws.com/Silver.zip
Gold - 1000 Images in dataset (3000 megabytes) https://pixels-data.s3.eu-west-2.amazonaws.com/Gold.zip