6DoF Pose estimation has been gaining increased importance in vision for over a decade, however it does not yet meet the reliability and accuracy standards for mass deployment in industrial robotics. To this effect, we present the Industrial Plenoptic Dataset (IPD): the first dataset for the co-evaluation of cameras, HDR, and algorithms targeted at reliable, high-accuracy industrial automation. Specifically, we capture 2,300 physical scenes of 20 industrial parts covering a 1m × 1m × 0.5m working volume, resulting in over 100,000 distinct object views. Each scene is captured with 13 well-calibrated multi-modal cameras including polarization and high-resolution structured light. In terms of lighting, we capture each scene at 4 exposures and in 3 challenging lighting conditions ranging from 100 lux to 100,000 lux. We also present, validate, and analyze robot consistency, an evaluation method targeted at scalable, high accuracy evaluation. We hope that vision systems that succeed on this dataset will have direct industry impact. The dataset and evaluation code are available at https://github.com/intrinsic-ai/ipd.

Rotation augmentations generally improve a model’s in-variance/equivariance to rotation - except in object detection. In object detection the shape is not known, therefore rotation creates a label ambiguity. We show that the de-facto method for bounding box label rotation, the Largest Box Method, creates very large labels, leading to poor performance and in many cases worse performance than using no rotation at all. We propose a new method of rotation augmentation that can be implemented in a few lines of code. First, we create a differentiable approximation of label accuracy and show that axis-aligning the bounding box around an ellipse is optimal. We then introduce Rotation Uncertainty (RU) Loss, allowing the model to adapt to the uncertainty of the labels. On five different datasets (including COCO, PascalVOC, and Transparent Object Bin Picking), this approach improves the rotational invariance of both one-stage and two-stage architectures when measured with AP, AP50, and AP75.

People often have to remove their phone from an inaccessible location like a pocket to view things like notifications and directions. We explore the idea of viewing such information through the fabric of a pocket using low resolution bright LED matrix displays. A survey confirms viewing information on inaccessible phones is desirable, and establishes types of pockets in garments worn by respondents and what objects are typically put in pockets. A technical evaluation validates that LED light can shine through many common garment fabrics. Based on these results, functional hardware prototypes are constructed to demonstrate different form factors of through-fabric display devices, such as a phone, wallet, a key fob, a pen, and earbud headphone case. A simple interaction vocabulary for viewing key information on these devices is described, and the social and technical aspects of the approach are discussed.