Single-Pixel Moving Object Classification with Differential Measuring in Transform Domain and Deep Learning

Abstract

Due to limited data transmission bandwidth and data storage space, it is challenging to perform fast-moving objects classification based on high-speed photography for a long duration. Here we propose a single-pixel classification method with deep learning for fast-moving objects. The scene image is modulated by orthogonal transform basis patterns, and the modulated light signal is detected by a single-pixel detector. Thanks to the property that the natural images are sparse in the orthogonal transform domain, we used a small number of basis patterns of discrete-sine-transform to obtain feature information for classification. The proposed neural network is designed to use single-pixel measurements as network input and trained by simulation single-pixel measurements based on the physics of the measuring scheme. Differential measuring can reduce the difference between simulation data and experiment data interfered by slowly varying noise. In order to improve the reliability of the classification results for fast-moving objects, we employed a measurement data rolling utilization approach for repeated classification. Long-duration classification of fast-moving handwritten digits that pass through the field of view successively is experimentally demonstrated, showing that the proposed method is superior to human vision in fast-moving digit classification. Our method enables a new way for fast-moving object classification and is expected to be widely implemented.

John von Neumann’s Space-Frequency Orthogonal Transforms

“…As already mentioned, the orthogonality of the 2D dictionary involves the orthogonality of the transform (12). Consequently, the practical inverse of JvNT can be computed like in [3].…”

“…Note, however, that the cosine coefficients are not necessarily symmetric. Recent publications, like [10][11][12][13], prove that the employment of these transforms still is quite intense.…”

Target recognition in Ghost Imaging from traditional to advance; a brief review

“…With a temporal resolution of 1.68 ms, the approach can effectively classify objects moving at up to 3.61 m/s in a 45 mm x 45 mm field of view [74]. For objects that move fast, a different approach utilizing deep learning for single-pixel classification for object recognition was put forth [75]. The suggested method, which is based on the structured detection scheme, modulates the object image using a limited number of discrete-sine-transform (DST) basis patterns to provide 1D single pixel measurements that are then submitted to a neural network for classification.…”

“…The technique demonstrates that it can classify moving objects in a noisy environment with great accuracy at a speed that is nearly impossible for human eye to realize. There's a new technique to recognize objects that move quickly with this proposed strategy [75]. A characteristics imaging model is built by X. Zou et al, [76] which is based on preprocessing in GI with DL in order to recognize target.…”

Single-Pixel Imaging Based on Deep Learning Enhanced Singular Value Decomposition

“…Jiang et al [39] proposed a novel SPI scheme for high-speed moving targets combined with a deep learning network and obtained reasonable reconstructions with a low sampling ratio of only 6%. Yao et al [40] proposed a single-pixel classification method with deep learning for fast-moving objects and obtained feature information for classification with S R = 3.8%. However, these data-driven networks suffer from problems such as generalizability and interpretability, which may prohibit their practical applications [41][42][43].…”