Cross-modality person re-identification (Re-ID) is more challenging than traditional visible Re-ID due to the huge cross-modality gap from heterogeneous images. To alleviate this problem, existing methods often utiliz...
详细信息
Cross-modality person re-identification (Re-ID) is more challenging than traditional visible Re-ID due to the huge cross-modality gap from heterogeneous images. To alleviate this problem, existing methods often utilize a dual path learning framework equipped with metric loss to learn discriminative features. Despite effectiveness, the inevitable degeneration of intra-modality discrimination by taking cross-modality discrimination into consideration is unsolvable. Such degeneration substantially hinders the model's capability of further improving feature representations. To mitigate this degeneration, we propose a Dual Mutual Learning (DML) method for cross-modality Re-ID which conducts mutual learning between the cross-modality and each of two single modalities. We design a triple-branch deep model containing the RGB and IR branches and the cross-modality branch. The cross-modality branch is designed to learn modality-invariant feature subspace for appearance similarity measurement. Both the RGB branch and IR branch provide attention supervision information to the cross-modality branch for attention feature alignment so as to enhance the intra-modality discrimination. Experimental results on two standard benchmarks demonstrate DML is superior to state-of-the-art methods.
Reconstructing the three-dimensional (3D) shape and texture of the face from a single image is a significant and challenging task in computer vision and graphics. In recent years, learning-based reconstruction methods...
详细信息
Blockchain transactions are increasingly vulnerable to sophisticated attacks, necessitating detection mechanisms that safeguard user privacy. Traditional methods, such as CTGAN and PATE-GAN, often sacrifice privacy to...
详细信息
Swarm situational awareness technology for uncrewed aerial vehicles (UAVs) is a core technology for mastering environmental information and gaining decision-making advantages. However, evaluation indicators and valida...
详细信息
Multimodal emotion recognition (MER) offers a more comprehensive and accurate understanding of human emotions by analyzing and integrating information from various modalities. However, previous work may have yet to fu...
详细信息
Head pose estimation methods can be generally classified into two categories: model-based and appearance-based methods. The model-based approach relies on facial landmarks for three-dimensional reconstruction, aiming ...
详细信息
We present LW-GeneFace, a lightweight and high-fidelity model for generalized audio-driven facial animation, in this paper. We develop this model by reducing the size while maintaining the synthetic quality of Ge...
详细信息
Aptamers are single-stranded DNA or RNA oligonucleotides that selectively bind to specific targets, making them valuable for drug design and diagnostic applications. Identifying the interactions between aptamers and t...
详细信息
Satellite-based and reanalysis precipitation products provide valuable information for various ***,their performance varies widely across regions due to different data sources and production *** paper evaluated the da...
详细信息
Satellite-based and reanalysis precipitation products provide valuable information for various ***,their performance varies widely across regions due to different data sources and production *** paper evaluated the daily performance of four precipitation products(MSWEP,ERA5,PERSIANN,and TRMM)for seven regions of the Chinese mainland,using observations from 2462 ground stations across the country as a *** used four statistical and four classification indicators to describe their spatial and temporal accuracy,and capability to detect precipitation events while analyzing their *** results show that according to the precipitation char-acteristics and accuracy of different types of precipitation products over the Chinese mainland,MSWEP was the most suitable product over the Chinese mainland,having the lowest root mean square error and mean absolute error,along with the highest coefficient of *** was followed by TRMM and ERA5,whereas PERSIANN lagged behind in terms of *** terms of different regions,MSWEP still performed well,especially in North China and East *** accuracy of the four precipitation products was relatively low in the summer months,and they all overestimated in the northwest *** other months,MSWEP and TRMM were better than PERSIANN and *** four precipitation products had good detection performance over the Chinese mainland,with probability of detection above ***,with the increase of precipitation threshold,the detection capability of the four products decreased,and MSWEP and ERA5 had good detection capability for moderate ***’s detection capability for heavy rain and rainstorms was better than that of the other three products,and PERSIANN’s detection capability for moderate rain,heavy rain and rainstorms was relatively poor,with a large deviation.
Atomic simulations aim to understand and predict complex physical phenomena,the success of which relies largely on the accuracy of the potential energy surface description and the efficiency to capture important rare ...
详细信息
Atomic simulations aim to understand and predict complex physical phenomena,the success of which relies largely on the accuracy of the potential energy surface description and the efficiency to capture important rare *** software(large-scale atomic simulation with a Neural Network Potential),released in 2018,incorporates the key ingredients to fulfill the ultimate goal of atomic simulations by combining advanced neural network potentials with efficient global optimization *** review introduces the recent development of the software along two main streams,namely,higher intelligence and more automation,to solve complex material and reaction *** latest version of LASP(LASP 3.7)features the global many-body function corrected neural network(G-MBNN)to improve the PES accuracy with low cost,which achieves a linear scaling efficiency for large-scale atomic *** key functionalities of LASP are updated to incorporate(i)the ASOP and ML-interface methods for finding complex surface and interface structures under grand canonic conditions;(ii)the ML-TS and MMLPS methods to identify the lowest energy reaction *** these powerful functionalities,LASP now serves as an intelligent data generator to create computational databases for end *** exemplify the recent LASP database construction in zeolite and the metal−ligand properties for a new catalyst design.
暂无评论