Model-based active learning to detect isometric deformable objects in the wild with deep architectures

TL;DR

This paper investigates the effectiveness of deep CNN-based object detection algorithms in recognizing isometric deformable objects under challenging imaging conditions, using synthetic and real data generated through rendering techniques.

Contribution

It introduces a novel approach combining rendering techniques with active learning to evaluate deep detection algorithms on difficult imaging scenarios for instance-level recognition.

Findings

Faster R-CNN achieved a mean Average Precision of 0.92 on synthetic images.

Detection algorithms perform well under various hard imaging conditions.

The study provides new insights into deep architectures' robustness for challenging object detection tasks.

Abstract

In the recent past, algorithms based on Convolutional Neural Networks (CNNs) have achieved significant milestones in object recognition. With large examples of each object class, standard datasets train well for inter-class variability. However, gathering sufficient data to train for a particular instance of an object within a class is impractical. Furthermore, quantitatively assessing the imaging conditions for each image in a given dataset is not feasible. By generating sufficient images with known imaging conditions, we study to what extent CNNs can cope with hard imaging conditions for instance-level recognition in an active learning regime. Leveraging powerful rendering techniques to achieve instance-level detection, we present results of training three state-of-the-art object detection algorithms namely, Fast R-CNN, Faster R-CNN and YOLO9000, for hard imaging conditions imposed into the scene by rendering. Our extensive experiments produce a mean Average Precision score of 0.92 on synthetic images and 0.83 on real images using the best performing Faster R-CNN. We show for the first time how well detection algorithms based on deep architectures fare for each hard imaging condition studied.