| Brain plays a decisive role in controlling human beings' thought,feeling and behavior,and closely relates to mental diseases.The basic unit of the brain is neurons.Structure determines function.Understanding brain structure from neuron level is of great value for understanding various functions of the brain and the etiology of mental illness.The Brainsmatics three-dimensional image data can help people understand the brain structure from the neuron level.With the development of imaging technology,the resolution of the imaging system is getting higher and higher,and the size of the imaging sample is getting bigger and bigger.The high resolution and large sample size determine that the three-dimensional image data obtained from the imaging system must be a massive amount of data.Huge amounts of data have caused great pressure on storage and transmission.It is of great significance to compress it effectively.The existing methods of three-dimensional image data compression is ineffective and inefficient,and it is difficult to apply to the data compression of massive three-dimensional microscopic images.Aiming at the above problems,a compression scheme of massive three-dimensional microscopic image data based on video coding is designed,and a compression tool suitable for massive three-dimensional microscopic image data is implemented.Focusing on the design of the scheme,we compare the performance of various video coders from the aspects of encoder's efficiency,compression ratio and compression time,and select the encoder suitable for massive three-dimensional microscopic image data.Based on x264 video encoder,a video compression scheme for 8-bits and 16-bits data is designed in this paper.And appropriate coding parameters for different data types has been selected.In video compression for 16-bits microscopic image data,a more reasonable compression scheme is proposed.At the same time,through the implementation of video coding interface,base on the existing blocking program,an easy to use video compression tool for massive three-dimensional image data and a matching format checking tool are implemented.To decompress the video compression data set,a data retrieval tool for ROI area of video compressed data is implemented.In the process of neuron reconstruction,through the use of video compressed data and by parallel acceleration,the speed of data acquisition operations involved in the reconstruction of neurons is speeded up notably,this improves the efficiency of neural reconstruction work.The compression tool implemented in this paper can compress the massive three-dimensional image data efficiently in a reasonable time.In the process of neuron reconstruction,by using video compression data and parallel acceleration,we significantly improve the speed of data acquisition operation in neuron reconstruction. |
PDF Full Text Request