SignMNIST

The original MNIST image dataset of handwritten digits is a popular benchmark for image-based machine learning methods but researchers have renewed efforts to update it and develop drop-in replacements that are more challenging for computer vision and original for real-world applications. The dataset format is patterned to match closely with the classic MNIST. Each training and test case represents a label (0-25) as a one-to-one map for each alphabetic letter A-Z (and no cases for 9=J or 25=Z because of gesture motions). The training data (27,455 cases) and test data (7172 cases) are approximately half the size of the standard MNIST but otherwise similar with a header row of label, pixel1, pixel2 ... pixel784 which represent a single 28x28 pixel image with grayscale values between 0-255. The train and test data have been concatenated and can be retrieved by selecting the first 27,455 rows for train and the last 7172 for test.