5 December 2019 in IEEE Transactions on Knowledge and Data Engineering
20 November 2014 in IEEE Transactions on Cybernetics
Object recognition and modeling have classically been studied separately, but practically, they are two closely correlated aspects. In this paper, by exploring the interrelations, we propose a framework to address these two problems at the same time, which we call simultaneous recognition and modeling. Differing from traditional recognition process which consists of off-line object model learning and on-line recognition procedures, our method is solely online. Starting with an empty object database, we incrementally build up object models while at the same time using these models to identify newly observed object views. In the proposed framework, objects are modeled as view graphs and a probabilistic observation model is presented. Both the appearance and the spatial structure of the object are examined, and a formulation based on maximum likelihood estimation is developed. Joint object recognition and modeling are achieved by solving the optimization problem. To evaluate the framework, we have developed a method for simultaneously learning multiple 3-D object models directly from the cluttered indoor environment and tested it using several everyday scenes. Experimental results demonstrate that the framework can cope with the recognition and modeling problem together nicely.
1 August 2016 in Computational Biology and Chemistry
20 March 2020 in IEEE Transactions on Medical Imaging
The goal of this work is to design high-resolution, high-contrast and robust MV adaptive beamforming algorithms, which are also implemented in real-time frame rate. Multi-operator optimization is introduced into MV adaptive beamforming in this work to propose a multi-operator MV adaptive beamforming algorithmic optimization framework. Based on the proposed algorithmic optimization framework, the algorithm optimization can be either conducted by activating a single optimization operator, or conducted by activating multiple optimization operators. The multi-operator MV (MOMV) adaptive beamforming algorithms are then derived from this framework. Moreover, in order to promote the real-time imaging capability of MOMV beamforming, a GPU-based parallel acceleration framework is proposed along with the algorithmic optimization framework by exploring the image-level coarse-grained parallelization and pixel-level fine-grained parallelization. GPU computing resource allocation strategy and memory access strategy are both explored to better design the acceleration framework. Comprehensive quantitative simulation evaluations and qualitative in vivo experiments of imaging performance are studied, and the results demonstrate that the proposed MOMV adaptive beamforming algorithms significantly improve the imaging performance as compared with other MV beamforming algorithms, which have high resolution, high contrast, good robustness, and real-time imaging capability with thousands of acceleration speedup. Furthermore, the MOMV beamforming algorithm without eigen-decomposition and projection optimization operator achieves much higher beamforming frame rate with little downgrade of image quality as compared with the MOMV beamforming algorithm with all optimization operators.
26 November 2019 in IEEE Access
With the rapid development of deep convolutional neural networks, more and more computer vision tasks have been well resolved. These convolutional neural network solutions rely heavily on the performance of the hardware. However, due to privacy issues or the network instability, we need to run convolutional neural networks on embedded platforms. Critical challenges will be raised by limited hardware resources on the embedded platform. In this paper, we design and implement an embedded inference framework to accelerate the inference of the convolutional neural network on the embedded platform. For this, we first analyzed the time-consuming layers in the inference process of the network, and then we design optimization methods for these layers. Also, we design a memory pool specifically for neural networks. Our experimental results show that our embedded inference framework can run a classification model MobileNet in 80ms and a detection model MobileNet-SSD in 155ms on Firefly-RK3399 development board.