China
Africa
Explore chapters and articles related to this topic
Algorithm/Architecture Coexploration
Published in Ling Guan, Yifeng He, Sun-Yuan Kung, Multimedia Image and Video Processing, 2012
Gwo Giun (Chris) Lee, He Yuan Lin, Sun Yuan Kung
In addition to the number of operations and data storages, the amount of data transfer is also an intrinsic complexity metric as executing an algorithm. In real-time applications, visual computing algorithms need to transfer a large amount of data within a specified time. Consequently, the average data transfer rate is a measure of the algorithmic complexity by providing an estimate of the amount of data transferred in 1 s. The average data transfer directly estimated from the algorithm is intrinsic to the algorithm itself and is independent of software or hardware implementations. As the design goes into the architectural level, the corresponding average bus or memory bandwidth can be estimated based on the algorithmic average data transfer rate. However, the peak bandwidth requirements of algorithms are significantly influenced by several architectural design details such as the memory hierarchy, data alignment in memory, data transaction type, and datapath. Therefore, the peak bandwidth is not intrinsic to algorithms but is dependent on architectures. Nevertheless, the peak bandwidth provides more insight into bus bit widths and clock speed in the ESL design.
Multiple factors influence coal and gangue image recognition method and experimental research based on deep learning
Published in International Journal of Coal Preparation and Utilization, 2023
Man Li, Xianli He, Yinxue Yuan, Maolin Yang
In coal mine production, sorting gangue from raw coal is a crucial link in coal preparation and also a critical step in the process of intellectualization of coal mines (Wang et al. 2020). In recent years, coal and gangue sorting robots have attracted wide attention in the industry. For coal and gangue sorting robots, accurate recognition and location of coal and gangue is the primary task to realize robot separation (Shang et al. 2022; Wang et al. 2020). To solve this problem, many researchers have studied the recognition and location methods of coal and gangue starting with image analysis and visual computing, and tried to apply them to practical production.