Image Recognition using Region Creep

TL;DR

This paper introduces 'Region Creep', a fast, auto-associative image classifier that segments images into regions, matches them efficiently, and uses neighborhood information to improve classification accuracy, achieving state-of-the-art results on handwritten digits.

Contribution

It presents a novel shallow architecture for image classification that leverages region-based matching and neighborhood aggregation, reducing memory issues and increasing speed.

Findings

Achieved state-of-the-art results on handwritten digit classification.

Demonstrated fast learning with a shallow architecture.

Showed that neighborhood information improves classification accuracy.

Abstract

This paper describes a new type of auto-associative image classifier that uses a shallow architecture with a very quick learning phase. The image is parsed into smaller areas and each area is saved directly for a region, along with the related output category. When a new image is presented, a direct match with each region is made and the best matching areas returned. Each area stores a list of the categories it belongs to, where there is a one-to-many relation between the input region and the output categories. The image classification process sums the category lists to return a preferred category for the whole image. These areas can overlap with each other and when moving from a region to its neighbours, there is likely to be only small changes in the area image part. It would therefore be possible to guess what the best image area is for one region by cumulating the results of its neighbours. This associative feature is being called 'Region Creep' and the cumulated region can be compared with train cases instead, when a suitable match is not found. Rules can be included and state that: if one set of pixels are present, another set should either be removed or should also be present, where this is across the whole image. The memory problems with a traditional auto-associative network may be less with this version and tests on a set of hand-written numbers have produced state-of-the-art results.