|
|
--- |
|
|
language: |
|
|
- en |
|
|
pretty_name: Binghamton HCI Sequence Digits Recognition Dataset |
|
|
size_categories: |
|
|
- 1K<n<10K |
|
|
--- |
|
|
|
|
|
# Dataset Card for Binghamton HCI Class Project 2 - Sequence Digits Recognition Task |
|
|
|
|
|
## Dataset Summary |
|
|
|
|
|
The **Binghamton HCI Class Project 2: Sequence Digits Recognition Task** dataset is a subset of the **MJ Synth Dataset**, specifically 1/10 of its original size. The MJSynth dataset consists of 9 million images covering 90,000 English words. For this project, the subset focuses exclusively on recognizing sequences of English digits (0-9) from images. |
|
|
|
|
|
The dataset retains the original training, validation, and test splits used in the full MJSynth dataset and is ideal for sequence modeling tasks, particularly multi-class classification of digit sequences. This dataset provides a smaller but diverse sample suitable for experimenting with models that handle digit sequence recognition, image-to-sequence tasks, and sequence-based classification. |
