2,189 Hits in 8.1 sec

Crossover Learning for Fast Online Video Instance Segmentation [article]

Shusheng Yang, Yuxin Fang, Xinggang Wang, Yu Li, Chen Fang, Ying Shan, Bin Feng, Wenyu Liu
2021 arXiv   pre-print
Modeling temporal visual context across frames is critical for video instance segmentation (VIS) and other video understanding tasks. In this paper, we propose a fast online VIS model named CrossVIS.  ...  By integrating with the instance segmentation loss, crossover learning enables efficient cross-frame instance-to-pixel relation learning and brings cost-free improvement during inference.  ...  video-based instance representation learning, and (2) the global balanced instance embedding branch for better online instance association.  ... 
arXiv:2104.05970v1 fatcat:ycw2a6ygwvgghkhcthvsbvywbm

From ideal to reality: segmentation, annotation, and recommendation, the vital trajectory of intelligent micro learning

Jiayin Lin, Geng Sun, Tingru Cui, Jun Shen, Dongming Xu, Ghassan Beydoun, Ping Yu, David Pritchard, Li Li, Shiping Chen
2019 World wide web (Bussum)  
Micro learning aims to effectively utilize learners' fragmented time to carry out personalized learning activities through online education resources.  ...  Our micro learning framework is firstly introduced in this paper from a higher perspective. Then we will review representative segmentation and annotation strategies in the e-learning domain.  ...  Extracting textual information for video learning material segmentation is a vital stage in many prior studies [10, 11, [13] [14] [15] .  ... 
doi:10.1007/s11280-019-00730-9 fatcat:7s5dus7tvze5rnzj7na2yszmdm

Multi-Adapter RGBT Tracking

Cheng Long Li, Andong Lu, Ai Hua Zheng, Zhengzheng Tu, Jin Tang
2019 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)  
for RGBT tracking.  ...  In this paper, we propose a novel Multi-Adapter convolutional Network (MANet) to jointly perform modality-shared, modality-specific and instance-aware feature learning in an end-to-end trained deep framework  ...  For the training of the instance adapter, we adopt offline and online algorithms to learn its parameters as shown in MDNet [24] .  ... 
doi:10.1109/iccvw.2019.00279 dblp:conf/iccvw/LiLZTT19 fatcat:sszwbdvrhvgznhkz2rtlhpcjhq

Deep Learning for Object Detection and Segmentation in Videos: Toward an Integration With Domain Knowledge

Athina Ilioudi, Azita Dabiri, Ben J. Wolf, Bart De Schutter
2022 IEEE Access  
Deep learning has enabled the rapid expansion of computer vision tasks from image frames to video segments.  ...  In order to address the existing challenges, a hybrid framework is proposed, where deep learning methods are coupled with domain knowledge.  ...  Crossover learning is integrated with the instance segmentation loss as an objective to obtain cross-frame instance segmentation consistency, achieving a low computational cost.  ... 
doi:10.1109/access.2022.3162827 fatcat:jdaz4brwgrhwrdbmzwqh7dg3wu

Assessing metacognitive knowledge in web-based CALL: a neural network approach

Shiou-Wen Yeh, Jia-Jiunn Lo
2005 Computers & Education  
The assessment of learnersÕ metacognitive knowledge level is crucial when developing computer-assisted language learning systems.  ...  The goal of this study is to present a neural network model that assesses automatically the learnerÕs metacognitive knowledge level by observing his/her online browsing behavior.  ...  Here, N is 20 and f i is the fitness value of the selected specie. (5) Crossover. There are three common crossover methods: one-point crossover, two-point crossover, and mask crossover.  ... 
doi:10.1016/j.compedu.2003.12.019 fatcat:x4ojxsuv4veyzaof3vewgghfva

RoadSpeedSense: Context-Aware Speed Profiling from Smart-phone Sensors

Ratna Mandal, Pallav Sonowal, Manish Kumar, Sujoy Saha, Subrata Nandi
2018 EAI Endorsed Transactions on Energy Web  
METHODS: We have used road surface quality, density of vehicles, type of neighborhood and road geometry for developing speed profile for a particular road segment.  ...  INTRODUCTION: There are several online mapping systems like Google Maps, Waze, Here, Apple Maps, Bing Maps, etc. which are developed to visualize real-time traffic conditions which rely on crowdsourced  ...  But, for the poor Intersection Density (Id) Road Intersection is a crossover of two or more than two road segments.  ... 
doi:10.4108/eai.7-1-2020.162802 fatcat:cy6pqi3cdfeuhlwlduzoqyuu7y

Online classification of visual tasks for industrial workflow monitoring

Athanasios Voulodimos, Dimitrios Kosmopoulos, Galina Veres, Helmut Grabner, Luc Van Gool, Theodora Varvarigou
2011 Neural Networks  
We propose a novel method to automatically segment the input stream and to classify the resulting segments using prior knowledge and hidden Markov models (HMMs), combined through a genetic algorithm.  ...  In this paper, we propose a framework for classification of visual tasks in industrial environments.  ...  ESNs offer several benefits, such as (i) fast and simple learning of many outputs simultaneously, (ii) the possibility of both offline and online learning, (iii) the capability of directly dealing with  ... 
doi:10.1016/j.neunet.2011.06.001 pmid:21757322 fatcat:c5z6mwaaijfaxcq3byxeem75c4

RelativeNAS: Relative Neural Architecture Search via Slow-Fast Learning [article]

Hao Tan, Ran Cheng, Shihua Huang, Cheng He, Changxiao Qiu, Fan Yang, Ping Luo
2021 arXiv   pre-print
As the key to efficient search, RelativeNAS performs joint learning between fast-learners (i.e. networks with relatively higher accuracy) and slow-learners in a pairwise manner.  ...  Moreover, since RelativeNAS only requires low-fidelity performance estimation to distinguish each pair of fast-learner and slow-learner, it saves certain computation costs for training the candidate architectures  ...  8M Video Classification Challenge, the first place in 2018 Drivable Area Segmentation Challenge for Autonomous Driving, 2011 HK PhD Fellow Award, and 2013 Microsoft Research Fellow Award (ten PhDs in  ... 
arXiv:2009.06193v3 fatcat:6ymzh27cpzgv5drgsemoybmdde

Deep learning in computer vision: A critical review of emerging techniques and application scenarios

Junyi Chai, Hao Zeng, Anming Li, Eric W.T. Ngai
2021 Machine Learning with Applications  
Deep learning has been overwhelmingly successful in computer vision (CV), natural language processing, and video/speech recognition. In this paper, our focus is on CV.  ...  was more efficient than learning layer parameters without referring to inputs.  ...  The learned feature is then adapted online for a given target video sequence through an adaptation module.  ... 
doi:10.1016/j.mlwa.2021.100134 fatcat:xwrvp237jnhqzoa3hdkj7qhanu

Implementation of Short Video Click-Through Rate Estimation Model Based on Cross-Media Collaborative Filtering Neural Network

Ying Feng, Guisheng Zhao, Gengxin Sun
2022 Computational Intelligence and Neuroscience  
In this paper, we analyze the construction of cross-media collaborative filtering neural network model to design an in-depth model for fast video click-through rate projection based on cross-media collaborative  ...  In this paper, by directly extracting the image features, behavioral features, and audio features of short videos as video feature representation, more video information is considered than other models  ...  fast video click-through rate estimation.  ... 
doi:10.1155/2022/4951912 pmid:35685157 pmcid:PMC9173947 fatcat:tq3tdoj5qzha3ltxvpaz7thge4

Multi-modal Visual Tracking: Review and Experimental Comparison [article]

Pengyu Zhang and Dong Wang and Huchuan Lu
2020 arXiv   pre-print
Finally, we discuss various future directions from different perspectives, including model design and dataset construction for further research.  ...  As animals move fast and irregularly, these trackers with online learning are easy to drift. While the target is in small size, CF can provide precise tracking results.  ...  Improved MDNet-based trackers achieve satisfactory performance on low-resolution, deformation, background clutter, fast motion, and thermal crossover.  ... 
arXiv:2012.04176v1 fatcat:pc3pt3hdavcp3pzij5sryvqe5y

Performance Comparisons of Bio-Micro Genetic Algorithms on Robot Locomotion

Francisco A. Chávez-Estrada, Jacobo Sandoval-Gutiérrez, Juan C. Herrera-Lozada, Mauricio Olguín-Carbajal, Daniel L. Martínez-Vázquez, Miguel Hernández-Bolaños, Israel Rivera-Zárate
2020 Applied Sciences  
The μ sGA algorithm uses less computing time since the individual is segmented into specific bytes.  ...  In contrast, the use of a computer and the high demand in computational resources for the GA are avoided.  ...  financial support for this research.  ... 
doi:10.3390/app10113863 fatcat:mzylp2c4sfglrgbsfehbqhe4bq

Extending Multi-Object Tracking systems to better exploit appearance and 3D information [article]

Kanchana Ranasinghe, Sahan Liyanaarachchi, Harsha Ranasinghe and Mayuka Jayawardhana
2019 arXiv   pre-print
We further explore heuristics based constraints for tracking in the Birds Eye View Space for efficiently exploiting 3D information as a constrained optimization problem for track prediction.  ...  This work involves exploiting temporally varying appearance and motion information for tracking.  ...  It has been trained on ImageNet datasets for similarity learning and can operate online.  ... 
arXiv:1912.11651v1 fatcat:sylnqh6a2jgdxibi4evyup5xt4


M. Pravin, Mrs. Marrynal S Eastaff
2021 Zenodo  
In this Paper we talked about on Major improvement of Neural organization models and its novel provisions of the particular Machine learning and Deep Learning Object recognition calculations in AI Software  ...  Subsequently, the AI programming learns all alone or with fractional help of human intelligence from examples and provisions in the large information.  ...  Networks for Accurate Object Detection and Segmentation.  ... 
doi:10.5281/zenodo.5674992 fatcat:i4ui2ghqhfeqzmiryp4lw4nqxa

Filter Learning from Deep Descriptors of a Fully Convolutional Siamese Network for Tracking in Videos

Hugo Chaves, Kevyn Ribeiro, André Brito, Hemerson Tacon, Marcelo Vieira, Augusto Cerqueira, Saulo Villela, Helena Maia, Darwin Concha, Helio Pedrini
2020 Proceedings of the 15th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications  
In recent years, the advancement of Deep Learning has revolutionized many areas in Computer Vision, including Visual Object Tracking.  ...  The improvements for the area under the curve metrics are 7.4% and 3.0%, for precision and success plots, respectively, being comparable to the state-of-the-art methods.  ...  The success in image classification inspired the development of solutions based on DL in many other CV applications as semantic segmentation, object localization, video analytics.  ... 
doi:10.5220/0008957606850694 dblp:conf/visapp/ChavesRBTVCVMCP20 fatcat:jrg3ytdimrh6dpq5rfizqrynpy
« Previous Showing results 1 — 15 out of 2,189 results