Multi-object tracking(MOT) aims to discover interesting foreground targets and meanwhile generate robust trajectories of these targets with different data association algorithms based on identity embedding features(Re...
详细信息
Multi-object tracking(MOT) aims to discover interesting foreground targets and meanwhile generate robust trajectories of these targets with different data association algorithms based on identity embedding features(Re-ID features) [1] or motion cues [2]. During the past few decades,some advanced MOT methods have achieved considerable progress due to their wide range of applications in modern autonomous driving and intelligent video analysis.
The International Telecommunication Union - Telecommunication Standardization Sector (ITU-T) Study Group 16 (SG16) meeting was held in Rennes, France, in April 2024 with online participation (hybrid). NTT Group made a...
详细信息
At NTT Human informatics Laboratories, we are engaged in Project Humanity aiming at solving problems on the basis of the human-centric principle of respecting the humanity that each person values and in a way that doe...
详细信息
This paper introduces distribution-flexible subset quantization(DFSQ), a post-training quantization method for super-resolution networks. Our motivation for developing DFSQ is based on the distinctive activation distr...
详细信息
This paper introduces distribution-flexible subset quantization(DFSQ), a post-training quantization method for super-resolution networks. Our motivation for developing DFSQ is based on the distinctive activation distributions of current super-resolution models, which exhibit significant variance across samples and channels. To address this issue, DFSQ conducts channel-wise normalization of the activations and applies distribution-flexible subset quantization(SQ), wherein the quantization points are selected from a universal set consisting of multi-word additive log-scale values. To expedite the selection of quantization points in SQ, we propose a fast quantization points selection strategy that uses K-means clustering to select the quantization points closest to the centroids. Compared to the common iterative exhaustive search algorithm, our strategy avoids the enumeration of all possible combinations in the universal set, reducing the time complexity from exponential to linear. Consequently, the constraint of time costs on the size of the universal set is greatly relaxed. Extensive evaluations of various super-resolution models show that DFSQ effectively improves performance even without fine-tuning. For example, for 4-bit EDSR×2 on the Urban benchmark, DFSQ obtains 0.242 dB PSNR gains.
Existing state-of-the-art methods for few-shot action recognition (FSAR) achieve promising performance by spatial and temporal modeling. However, most current methods ignore the importance of edge information and moti...
详细信息
To advance research on more intuitive interfaces that connect people to people and people to machines and to achieve interfaces that are user-friendly for everyone regardless of age or disability, we focus on research...
详细信息
Memory deduplication is widely used to effectively reduce the memory footprint in operating systems, while huge pages are employed to enhance access performance by increasing TLB hit rates. Unfortunately, duplicate hu...
详细信息
Unmanned aerial vehicles(UAVs)have opened up various applications with significant socioeconomic benefits,but require reliable connectivity for control and data *** situations where obstacles separate aerial vehicles ...
详细信息
Unmanned aerial vehicles(UAVs)have opened up various applications with significant socioeconomic benefits,but require reliable connectivity for control and data *** situations where obstacles separate aerial vehicles and ground operators,the cellular network can be used to establish beyond the visual line of sight(BVLOS)communication ***,the cellular network is designed to optimize terrestrial user performance,so investigating its connectivity performance for aerial users is *** paper presents an investigation into the upper-bound throughput estimations of the cellular network based on practical RSS mea-surements for both aerial and terrestrial users occupying the network resources simultaneously,in a typical sub-urban environment in *** measurements of received signal strength(RSS)were taken to quantify the throughput performance for each category of *** indicate that the cellular network is feasible as a communication link for UAVs at altitudes below 40 ***,the overall throughput performance degrades when aerial users are present,and the probability of aerial user arrival must be kept at a minimum level beyond this *** this,the mean aerial throughput outperforms the terrestrial throughput below 40 m.
Virtual Reality as a stimulus to evoke emotion in emotion recognition studies has been emerging in recent years, the objective of this research is to do a comparative analysis between classifiers such as Deep learning...
详细信息
"Cho Kabuki 2022 Powered by NTT," a kabuki play sponsored by Shochiku Co., Ltd., is the first social implementation of Another Me, a technology for creating a human digital twin that reproduces the appearanc...
详细信息
暂无评论