Retrieving an occluded pedestrian remains a challenging problem in person re-identification (re-id). Most existing methods utilize external detectors to disentangle the visible body parts. However, these are unstable due domain bias and consume numerous computing resources. In this paper, we propose novel lightweight Part-based Representation Enhancement (PRE) network for re-id that takes full advantages of local correlations aggregate distinctive information features without relying on...

Introduction Grapes are prone to various diseases throughout their growth cycle, and the failure promptly control these can result in reduced production even complete crop failure. Therefore, effective disease is essential for maximizing grape yield. Accurate identification plays a crucial role this process. In paper, we proposed real-time lightweight detection model called Fusion Transformer YOLO 4 detection. The primary source of dataset comprises RGB images acquired from plantations...

To tackle the high resource consumption in occluded person re-identification, sparse attention mechanisms based on Vision Transformers (ViTs) have become popular. However, they often suffer from performance degradation with long sequences, omission of crucial information, and token representation convergence. address these issues, we introduce AIRHF-Net: an Adaptive Interaction Representation Hierarchical Fusion Network, named AIRHF-Net, designed to enhance pedestrian identity recognition...

Fingerprint classification is an important indexing scheme to reduce fingerprint matching time for a large database efficient large-scale identification. The abilities of Curvelet transform capturing directional edges images make the suitable be classified higher accuracy. This paper presents algorithm combining (CT) and gray-level cooccurrence matrix (GLCM). Firstly, we use fast discrete warping (FDCT_WARPING) decompose original image into five scales coefficients construct filter by...

Because the corn vein and noise influence contour extraction of maize leaf disease, we put forward a new recognition algorithm based on Curvelet Shape Context (SC). This method can improve speed accuracy disease recognition. Firstly, use Seeded Regional Growing (SRG) to segment image. Secondly, Modulus Correlation (CMC) is extract effective disease. Thirdly, combine CMC with SC obtain histogram features then these calculate similarities between template image target Finally, adopt n -fold...

Feature fusion is widely used in person re-identification (re-ID) and has been proven effective. However, it difficult to know which features are effective identify a specific how fuse explore complementary information apply the advantages of each feature. Motivated by these problems, this paper proposes new method re-ID recognition results multiple at rank level. Three innovations included method: first, metric spaces constructed based on correlation different generate results; second, most...

Pedestrian re-identification (re-ID) is a video surveillance technology for specific pedestrians in non-overlapping multi-camera scenes. However, due to the influence of dramatic changes perspectives and pedestrian occasions, it still huge challenge find stable, reliable algorithm high accuracy rate. In this paper, multiple layers re-ranking approach proposed jointly account that challenge. The re-ID viewed as metrics ranking optimizing problem by using Multiple Layers Re-ranking framework....

Aiming at the defect of AKAZE algorithm with low feature matching accuracy, this paper proposes an improved and I-SURF to optimize matching. First, composite conduction function according scale factors is used nonlinear diffusion filtering. Then SURF descriptors are calculated by second order derivative multi-scale template obtain more high-quality point pairs, Best Bin First (BBF) method high threshold adopted improve number pairs during Finally, MSAC further remove redundant points. The...

Based on the theoretical derivation of one-dimensional fractal, a new method to realize on-line tool wear monitoring micro end is proposed. First, several different conditions are chosen as comparison samples, which includes rate 0, 10, 20, 30, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$i$</tex> and tipping respectively. The vibration signals all samples collected in time domain, range xmlns:xlink="http://www.w3.org/1999/xlink">$k$</tex>...

Automated fine breeding is an important means to improve the scientific raising level, pork yield and economic benefit. At present, intelligent automatic feeding environmental control equipment have been basically popularized in most pig farms. However, judging whether behavior abnormal, it mainly depends on intuition experience of breeder. This method not only consumes a lot time energy, accuracy high. Based deep learning theory. Firstly, abnormal data homogeneous sensors are removed...