Triple-Sigmoid Activation Function for Deep Open-Set Recognition
Sign Up to like & getrecommendations! Published in 2022 at "IEEE Access"
DOI: 10.1109/access.2022.3192621
Abstract: Traditional models for various machine learning problems such as image classification perform well only under the assumption of a closed set. This implies that inputs must belong to the classes for which the models were… read more here.
Keywords: triple sigmoid; activation function; model; set recognition ... See more keywords
A Research of Gas Open-Set Identification Based on Data Augmentation Algorithm
Sign Up to like & getrecommendations! Published in 2023 at "IEEE Access"
DOI: 10.1109/access.2023.3247571
Abstract: Significant progress has been made in convolutional neural networks (CNN) based gas recognition. However, existing electronic nose (e-Nose) algorithms all use the closed-set assumption that the test and training samples are in the same label… read more here.
Keywords: gas; set recognition; set identification; open set ... See more keywords
An Open Set Recognition Method for SAR Targets Based on Multitask Learning
Sign Up to like & getrecommendations! Published in 2022 at "IEEE Geoscience and Remote Sensing Letters"
DOI: 10.1109/lgrs.2021.3079418
Abstract: Most of the existing synthetic aperture radar (SAR) automatic target recognition (ATR) methods aim at the closed set situation, in which the classes of targets in the test set have appeared in the training set.… read more here.
Keywords: recognition; multitask learning; based multitask; method ... See more keywords
Extreme Value Meta-Learning for Few-Shot Open-Set Recognition of Hyperspectral Images
Sign Up to like & getrecommendations! Published in 2023 at "IEEE Transactions on Geoscience and Remote Sensing"
DOI: 10.1109/tgrs.2023.3276952
Abstract: Recent advancements in prototype-based few-shot open-set recognition (FSOSR) approaches reject outliers based on the high metric distances from the known class prototypes and fail to distinguish spectrally fine-grained land cover outliers. Learning only the Euclidean… read more here.
Keywords: known class; shot open; class; meta ... See more keywords
Open Set Recognition of Timber Species Using Deep Learning for Embedded Systems
Sign Up to like & getrecommendations! Published in 2019 at "IEEE Latin America Transactions"
DOI: 10.1109/tla.2019.9011545
Abstract: Reliable and rapid identification of timber species is a very relevant issue for many countries in South America and especially for Peru, which is the second country with the largest extent of tropical forest, and… read more here.
Keywords: recognition timber; open set; set recognition; timber species ... See more keywords
Measuring Human Perception to Improve Open Set Recognition
Sign Up to like & getrecommendations! Published in 2022 at "IEEE transactions on pattern analysis and machine intelligence"
DOI: 10.48550/arxiv.2209.03519
Abstract: The human ability to recognize when an object belongs or does not belong to a particular vision task outperforms all open set recognition algorithms. Human perception as measured by the methods and procedures of visual… read more here.
Keywords: recognition; reaction time; human perception; set recognition ... See more keywords