Back to datasets
Dataset assetOpen Source CommunityImage RecognitionMulti‑View Data
MNIST Multiview Datasets
The MNIST Multiview Datasets comprise two four‑view datasets, each view being an R<sup>14 × 14</sup> vector. MNIST<sub>1</sub> generates four views by dividing the image into quadrants. MNIST<sub>2</sub> generates four overlapping views around the image centre, introducing redundancy between views.
Source
github
Created
Jan 19, 2019
Updated
May 8, 2024
Signals
166 views
Availability
Linked source ready
Overview
Dataset description and usage context
MNIST Multiview Datasets Overview
Dataset Description
The MNIST Multiview Datasets contain two sub‑datasets:
- MNIST1: Four views generated by splitting the original image into quadrants. Each view is a vector of size R14 × 14.
- MNIST2: Four overlapping views generated around the image centre, introducing redundancy among views.
Dataset Structure
- MNIST_1 folder: Contains 4 files corresponding to the 4 views.
- MNIST_2 folder: Also contains 4 files corresponding to the 4 views.
Each file inside the folders corresponds to the view data of the respective sub‑dataset.
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.