You Better Look Twice: a new perspective for designing accurate detectors with reduced computations

TL;DR

BLT-net is a novel two-stage object detection architecture that significantly reduces computational costs by efficiently separating background from objects and dynamically adjusting processing resolution, especially effective for scenes with varied object sizes.

Contribution

Introduces BLT-net, a low-computation, two-stage detection architecture that improves efficiency by separating background and dynamically adjusting resolution based on object scale.

Findings

Reduces computations by a factor of 4-7 on pedestrian datasets.

Maintains high detection accuracy with minimal degradation.

Applicable to scenes with high background and varied object sizes.

Abstract

General object detectors use powerful backbones that uniformly extract features from images for enabling detection of a vast amount of object types. However, utilization of such backbones in object detection applications developed for specific object types can unnecessarily over-process an extensive amount of background. In addition, they are agnostic to object scales, thus redundantly process all image regions at the same resolution. In this work we introduce BLT-net, a new low-computation two-stage object detection architecture designed to process images with a significant amount of background and objects of variate scales. BLT-net reduces computations by separating objects from background using a very lite first-stage. BLT-net then efficiently merges obtained proposals to further decrease processed background and then dynamically reduces their resolution to minimize computations. Resulting image proposals are then processed in the second-stage by a highly accurate model. We demonstrate our architecture on the pedestrian detection problem, where objects are of different sizes, images are of high resolution and object detection is required to run in real-time. We show that our design reduces computations by a factor of x4-x7 on the Citypersons and Caltech datasets with respect to leading pedestrian detectors, on account of a small accuracy degradation. This method can be applied on other object detection applications in scenes with a considerable amount of background and variate object sizes to reduce computations.