Filters








874 Hits in 4.3 sec

Solving Long-tailed Recognition with Deep Realistic Taxonomic Classifier [article]

Tz-Ying Wu, Pedro Morgado, Pei Wang, Chih-Hui Ho, Nuno Vasconcelos
2020 arXiv   pre-print
Motivated by this, a deep realistic taxonomic classifier (Deep-RTC) is proposed as a new solution to the long-tail problem, combining realism with hierarchical predictions.  ...  Long-tail recognition tackles the natural non-uniformly distributed data in real-world scenarios.  ...  Long-tailed Recognition Existing approaches formulate long-tailed recognition as flat classification, solved by some variant of the softmax classifier.  ... 
arXiv:2007.09898v1 fatcat:tttr66jedfgwrom46ymt4vzofq

Deep Long-Tailed Learning: A Survey [article]

Yifan Zhang, Bingyi Kang, Bryan Hooi, Shuicheng Yan, Jiashi Feng
2021 arXiv   pre-print
Deep long-tailed learning, one of the most challenging problems in visual recognition, aims to train well-performing deep models from a large number of images that follow a long-tailed class distribution  ...  However, long-tailed class imbalance, a common problem in practical visual recognition tasks, often limits the practicality of deep network based recognition models in real-world applications, since they  ...  Realistic taxonomic classifier (RTC) [85] proposed to address class imbalance with hierarchical classification.  ... 
arXiv:2110.04596v1 fatcat:lpvt2x6cv5crxm2qxdctjrlkqq

IP102: A Large-Scale Benchmark Dataset for Insect Pest Recognition

Xiaoping Wu, Chi Zhan, Yu-Kun Lai, Ming-Ming Cheng, Jufeng Yang
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Specifically, it contains more than 75, 000 images belonging to 102 categories, which exhibit a natural long-tailed distribution.  ...  In addition, we annotate about 19, 000 images with bounding boxes for object detection.  ...  Natural Science Foundation of Tianjin, China (No. 18JCYBJC15400, 18ZXZNGX00110, 17JCJQJC43700), the National Youth Talent Support Program, and the Open Project Program of the National Laboratory of Pattern Recognition  ... 
doi:10.1109/cvpr.2019.00899 dblp:conf/cvpr/WuZLCY19 fatcat:ixndm7xv3vaaxojr7tp22uuixq

Categorizing plant images at the variety level: Did you say fine-grained?

Julien Champ, Titouan Lorieul, Pierre Bonnet, Najate Maghnaoui, Christophe Sereno, Thierry Dessup, Jean-Michel Boursiquot, Laurent Audeguin, Thierry Lacombe, Alexis Joly
2016 Pattern Recognition Letters  
lying in the long tail of data occurrences (i.e classes with very few or none training samples available).  ...  The GoogLeNet architecture consists of a 22 layers deep network with a softmax loss as the classifier on top, and is composed of three "inception modules" stacked on top of each other.  ... 
doi:10.1016/j.patrec.2016.05.022 fatcat:fz5x7yzetrfbnnubwvdmlfi2xq

Deep learning and computer vision will transform entomology

Toke T. Høye, Johanna Ärje, Kim Bjerge, Oskar L. P. Hansen, Alexandros Iosifidis, Florian Leese, Hjalte M. R. Mann, Kristian Meissner, Claus Melvad, Jenni Raitoharju
2021 Proceedings of the National Academy of Sciences of the United States of America  
We identify four focal areas, which will facilitate this transformation: 1) validation of image-based taxonomic identification; 2) generation of sufficient training data; 3) development of public, curated  ...  reference databases; and 4) solutions to integrate deep learning and molecular tools.  ...  These approaches can help solve long standing challenges in ecology and biodiversity research and also address pressing issues in insect population monitoring (32, 33) .  ... 
doi:10.1073/pnas.2002545117 pmid:33431561 fatcat:3m4xtz5365awlpv6nlmunkuwl4

Fine-Grained Image Analysis with Deep Learning: A Survey [article]

Xiu-Shen Wei and Yi-Zhe Song and Oisin Mac Aodha and Jianxin Wu and Yuxin Peng and Jinhui Tang and Jian Yang and Serge Belongie
2021 arXiv   pre-print
Capitalizing on advances in deep learning, in recent years we have witnessed remarkable progress in deep learning powered FGIA.  ...  Fine-grained image analysis (FGIA) is a longstanding and fundamental problem in computer vision and pattern recognition, and underpins a diverse set of real-world applications.  ...  Novel properties of these datasets include the fact that they are large-scale, have a hierarchical structure, exhibit a domain gap, and form a long-tailed distribution.  ... 
arXiv:2111.06119v2 fatcat:ninawxsjtnf4lndtqquuwl3weq

Image Classification with Small Datasets: Overview and Benchmark

Lorenzo Brigato, Bjorn Barz, Luca Iocchi, Joachim Denzler
2022 IEEE Access  
Image classification with small datasets has been an active research area in the recent past.  ...  Indeed, only a single specialized method dating back to 2019 clearly wins our benchmark and outperforms the baseline classifier.  ...  Long-tailed recognition, also known as unbalanced classification, is a largely researched area, as high-class imbalance naturally occurs in many classification problems [25] .  ... 
doi:10.1109/access.2022.3172939 fatcat:f74jdgenanhfragbabhuuiaww4

Open-world Machine Learning: Applications, Challenges, and Opportunities [article]

Jitendra Parmar, Satyendra Singh Chouhan, Vaskar Raychoudhury, Santosh Singh Rathore
2022 arXiv   pre-print
Whereas open-world machine learning (OWML) deals with unseen classes. In this paper, first, we present an overview of OWML with importance to the real-world context.  ...  The framework is an end to end OWR (open-world recognition). To detect the instances from unknown classes they proposed Multi-stage Deep Classifier Cascades (MDCC).  ...  To achieve more realistic results the system must deal with both empirical and open-space risk.  ... 
arXiv:2105.13448v2 fatcat:rv6f42sdvvajnhub4uguuhb2cy

Commonsense Reasoning, Commonsense Knowledge, and The SP Theory of Intelligence [article]

J Gerard Wolff
2018 arXiv   pre-print
A solution is presented to the 'long tail' problem described by DM. The SP system has some potentially useful things to say about several of DM's objectives for research in CSR and CSK.  ...  In four main sections, the paper describes: 1) The main problems to be solved; 2) Other research on CSR and CSK; 3) Why the SP system may prove useful with CSR and CSK 4) How examples described by DM may  ...  with a solution of the 'long tail' problem.  ... 
arXiv:1609.07772v2 fatcat:qgid34xsbrec7omjtg5kbaxlqu

Open-world Machine Learning: Applications, Challenges, and Opportunities

Jitendra Parmar, Satyendra Singh Chouhan, Vaskar Raychoudhury, Santosh S. Rathore
2022 ACM Computing Surveys  
Open-world Machine Learning (OWML) is a novel technique, which deals with unseen classes.  ...  Finally, the paper concludes with a set of research gaps, open challenges, and future directions.  ...  The Supplementary ile is uploaded with the paper and available on the GitHub repository. 2  ... 
doi:10.1145/3561381 fatcat:hf4ownhjqjdfrgh6g3sanzz2ge

Challenges and Advances in the Taxonomy of Deep-Sea Peracarida: From Traditional to Modern Methods

Inmaculada Frutos, Stefanie Kaiser, Łukasz Pułaski, Łukasz Pułaski, Maciej Studzian, Maciej Studzian, Magdalena Błażewicz
2022 Frontiers in Marine Science  
In the deep sea, particular challenges are posed to the taxonomic differentiation of species.  ...  The taxonomic impediment, i.e. the shortage of taxonomists in view of a high undescribed biodiversity, is discussed in the context of the existing large taxonomic knowledge gaps in connection with the  ...  While wellestablished traditional methods are often still in use to describe and classify deep-sea Peracarida, new methodologies, notably molecular and microscopic imaging tools, have taken their taxonomic  ... 
doi:10.3389/fmars.2022.799191 doaj:14e36667e46141419df5b46d97c501eb fatcat:7cng4rpl4jdarisvn4pb3jhasq

A Survey of Deep Learning-based Object Detection

Licheng Jiao, Fan Zhang, Fang Liu, Shuyuan Yang, Lingling Li, Zhixi Feng, Rong Qu
2019 IEEE Access  
With the rapid development of deep learning networks for detection tasks, the performance of object detectors has been greatly improved.  ...  Object detection is one of the most important and challenging branches of computer vision, which has been widely applied in peoples life, such as monitoring security, autonomous driving and so on, with  ...  [152] present a deep learning framework on synthetic aperture radar (SAR) automatic target recognition. Long et al. [153] concentrate on automatically and accurately locating objects.  ... 
doi:10.1109/access.2019.2939201 fatcat:jesz2av2tjbkxfpaqyecptgls4

LifeCLEF 2014: Multimedia Life Species Identification Challenges [chapter]

Alexis Joly, Hervé Goëau, Hervé Glotin, Concetto Spampinato, Pierre Bonnet, Willem-Pier Vellinga, Robert Planque, Andreas Rauber, Robert Fisher, Henning Müller
2014 Lecture Notes in Computer Science  
Each task is based on large and real-world data and the measured challenges are defined in collaboration with biologists and environmental stakeholders in order to reflect realistic usage scenarios.  ...  Using multimedia identification tools is considered as one of the most promising solution to help bridging the taxonomic gap and build accurate knowledge of the identity, the geographic distribution and  ...  One of the key challenge is notably to deal with the long tail of species that are represented with much less images than the top-500 most common species that we targeted within BirdCLEF and PlantCLEF  ... 
doi:10.1007/978-3-319-11382-1_20 fatcat:oanew4xgx5h2tcknr72hoiwqry

Mid-level Representation for Visual Recognition [article]

Moin Nabi
2015 arXiv   pre-print
In the case of image understanding, we focus on object detection/recognition task.  ...  Employing mid-level representation, in particular, shifted the paradigm in visual recognition.  ...  Long-tail Distribusion of Subcategories: A recent argument in the object recognition community is that the frequency of samples in subcategories of an object category follow a long-tail distribution: a  ... 
arXiv:1512.07314v1 fatcat:knmhkwxqk5aczis7ce6g2sv2wm

Crowdsourcing Biodiversity Monitoring

Alexis Joly, Hervé Goëau, Julien Champ, Samuel Dufour-Kowalski, Henning Müller, Pierre Bonnet
2016 Proceedings of the 2016 ACM on Multimedia Conference - MM '16  
With the recent advances in computer vision, we see the emergence of more and more effective identification tools allowing to set-up largescale data collection platforms such as the popular Pl@ntNet initiative  ...  However, even with such big data, sparsity would still be a challenge, in particular for the vast majority of species lying in the long tail distribution.  ...  Overall, the proposed metric makes it possible to compensate the long-tail distribution effects of social data.  ... 
doi:10.1145/2964284.2976762 dblp:conf/mm/JolyGCDMB16 fatcat:ikwvk7dgpfd5jieug7jtjz5fd4
« Previous Showing results 1 — 15 out of 874 results