A knowledge-embedded lossless image compressing method for high-throughput corrosion experiment

High-throughput experiment refers to carry out a large number of tests and attain various characterizations in one experiment with highly integrated sample or facility, widely adopted in biology, medicine, and materials areas. Consequently, the storing and treating of data bring new challenges because of large amount of real-time data, especially high-resolution images. To improve the storing and treating efficiency of high-throughput image, a knowledge-embedded lossless image compressing method is proposed. Based on the similarity of a series of high-throughput images, it accomplishes the high compression ratio according to the difference between the target images and one reference image. Meanwhile, the knowledge extracted from the image, such as edge information and differences from the reference image, is recorded into the compressed file. The key steps include similarity comparison, edge detection, coordinate transformation, and dictionary encoding. The method has been successfully applied into high-throughput corrosion experiment facility, a typical intelligent cyber-physical system. To evaluate the performance, corrosion metal, face, and flower images are compressed by our method and other lossless image compression methods. The results show that our method has fairly high compression ratio. Moreover, the embedded knowledge can be read directly from the compressed file to support further study.

Solder joint imagery compressing and recovery based on compressive sensing

Purpose – The purpose of this paper is to develop an improved compressive sensing algorithm for solder joint imagery compressing and recovery. The improved algorithm can improve the performance in terms of peak signal to noise ratio (PSNR) of solder joint imagery recovery. Design/methodology/approach – Unlike the traditional method, at first, the image was transformed into a sparse signal by discrete cosine transform; then the solder joint image was divided into blocks, and each image block was transformed into a one-dimensional data vector. At last, a block compressive sampling matching pursuit was proposed, and the proposed algorithm with different block sizes was used in recovering the solder joint imagery. Findings – The experiments showed that the proposed algorithm could achieve the best results on PSNR when compared to other methods such as the orthogonal matching pursuit algorithm, greedy basis pursuit algorithm, subspace pursuit algorithm and compressive sampling matching pursuit algorithm. When the block size was 16 × 16, the proposed algorithm could obtain better results than when the block size was 8 × 8 and 4 × 4. Practical implications – The paper provides a methodology for solder joint imagery compressing and recovery, and the proposed algorithm can also be used in other image compressing and recovery applications. Originality/value – According to the compressed sensing (CS) theory, a sparse or compressible signal can be represented by a fewer number of bases than those required by the Nyquist theorem. The findings provide fundamental guidelines to improve performance in image compressing and recovery based on compressive sensing.

Wavelet Transforms of Image Reconstruction Based on Compressed Sampling

A compressive sensing technique for image signal to cope with image compression and restoration is adopted in this paper. First of all wavelet transforms method is applied in image compressing to preserve the constructive, Secondly, sparse matrix is available by required wavelet ratio. Thirdly, the compressing image is used to restoration the original image. Experimental results show that the proposed algorithm is effective and compares favorably with existing techniques.