Learn More
Deep learning has become a ubiquitous technology to improve machine intelligence. However, most of the existing deep models are structurally very complex, making them difficult to be deployed on the mobile platforms with limited computational power. In this paper, we propose a novel network compression method called dynamic network surgery, which can(More)
This paper presents incremental network quantization (INQ), a novel method, targeting to efficiently convert any pre-trained full-precision convolutional neural network (CNN) model into a low-precision version whose weights are constrained to be either powers of two or zero. Unlike existing methods which are struggled in noticeable accuracy loss, our INQ(More)
Scale invariant feature transform (SIFT) is an approach for extracting distinctive invariant features from images, and it has been successfully applied to many computer vision problems (e.g. face recognition and object detection). However, the SIFT feature extraction is compute-intensive, and a real-time or even super-real-time processing capability is(More)
Almost all of the current top-performing object detection networks employ region proposals to guide the search for object instances. State-of-the-art region proposal methods usually need several thousand proposals to get high recall, thus hurting the detection efficiency. Although the latest Region Proposal Network method gets promising detection accuracy(More)
This paper parallelizes and characterizes an important computer vision application — Scale Invariant Feature Transform (SIFT) both on a Symmetric Multiprocessor (SMP) platform and a large scale Chip Multiprocessor (CMP) simulator. SIFT is an approach for extracting distinctive invariant features from images and has been widely applied. In many computer(More)
Depth-map merging is one typical technique category for multi-view stereo (MVS) reconstruction. To guarantee accuracy, existing algorithms usually require either sub-pixel level stereo matching precision or continuous depth-map estimation. The merging of inaccurate depth-maps remains a challenging problem. This paper introduces a bundle optimization method(More)
The Emotion Recognition in the Wild (EmotiW) Challenge has been held for three years. Previous winner teams primarily focus on designing specific deep neural networks or fusing diverse hand-crafted and deep convolutional features. They all neglect to explore the significance of the latent relations among changing features resulted from facial muscle(More)
Forty-four flax genotypes with a diverse genetic background were evaluated for anther culture response using a standard anther culture protocol in order to determine the feasibility to initiate a routine haploid production system in applied breeding programs. A strong genotype effect on callus induction and shoot regeneration in anther culture was found in(More)
Anther culture is considered as the most successful method of producing doubled haploid plants in flax. The efficiency of shoot regeneration from anther culture has been improved dramatically by optimizing culture temperature and callus induction medium. However, shoot elongation has become increasingly the limiting factor for further improvement of the(More)
This paper investigates the problem of selecting an energy-efficient transmission power for ad hoc networks. Specifically, we examine how to determine a network-wide transmission power that minimizes global energy consumption, with desired lower bounds on capacity and connectivity. Based on recent empirical measurements of energy consumption in wireless(More)