Published Paper Details:

This paper presents a novel approach to object detection using deep convolutional neural networks and Big Data. The aim is to build an accurate, fast and reliable object detection system, which is a vital element of an autonomous robotic vision platform; it is a key element for object estimation and automated event based systems. Recent work in deep neural networks has led to the development of a state-of-the-art object detector termed Faster Region-based CNN (Mask R-CNN). This model, through transfer learning, for the task of object detection using imagery obtained from two modalities: color (RGB) and Near-Infrared (NIR). Early and late fusion methods are explored for combining the multi-modal (RGB and NIR) information. This leads to a novel multi-modal masked R-CNN model, which achieves state-of-the-art results compared to prior work with the F1 score, which takes into account both precision and recall performances improving from 0.781 to 0.812 for the detection of solid metal objects. In addition to improved accuracy, this approach is also much quicker to deploy for new objects, as it requires bounding box annotation rather than pixel-level annotation (annotating bounding boxes is approximately an order of magnitude quicker to perform). The model is retrained to perform the detection of four types of metal objects, with the entire process taking four hours to annotate and train the new model per solid block. A key benefit of Deep Learning is the analysis and learning of massive amounts of unsupervised data, making it a valuable tool for Big Data Analytics where raw data is largely unlabeled and un-categorized. In the present study, we explore how Deep Learning can be utilized for addressing some important problems in Big Data Analytics, including extracting complex patterns from massive volumes of data, semantic indexing, data tagging, fast information retrieval, and simplifying discriminative tasks.