Filters








192 Hits in 5.9 sec

HYPE: A Benchmark for Human eYe Perceptual Evaluation of Generative Models [article]

Sharon Zhou, Mitchell L. Gordon, Ranjay Krishna, Austin Narcomey, Li Fei-Fei, Michael S. Bernstein
2019 arXiv   pre-print
We construct Human eYe Perceptual Evaluation (HYPE) a human benchmark that is (1) grounded in psychophysics research in perception, (2) reliable across different sets of randomly sampled outputs from a  ...  Generative models often use human evaluations to measure the perceived quality of their outputs.  ...  We also thank Durim Morina and Gabby Wright for their contributions to the HYPE system and website. M.L.G. was supported by a Junglee Corporation Stanford Graduate Fellowship.  ... 
arXiv:1904.01121v4 fatcat:j2kvqxry3rcx5p7uppvb2gzd3a

Pros and Cons of GAN Evaluation Measures: New Developments [article]

Ali Borji
2021 arXiv   pre-print
With the dramatic progress in generative modeling, a suite of new quantitative and qualitative techniques to evaluate models has emerged.  ...  This work is an update of a previous paper on the same topic published a few years ago.  ...  These measures typically focus on how convincing a generated image is from human perception perspective. Human Eye Perceptual Evaluation (HYPE) Zhou et al.  ... 
arXiv:2103.09396v3 fatcat:shsmleblqfgzfpa3w5mwt7xyma

RainNet: A Large-Scale Dataset for Spatial Precipitation Downscaling [article]

Xuanhong Chen, Kairui Feng, Naiyuan Liu, Yifan Lu, Zhengyan Tong, Bingbing Ni, Ziang Liu, Ning Lin
2020 arXiv   pre-print
Deep learning is a promising potential solution for downscaling.  ...  Eight metrics specifically considering the physical property of the data set are raised, while fourteen models are evaluated on the proposed dataset.  ...  We can see that VGG19 highlights the hurricane eye, which illustrates that the perceptual loss is also effective for the precipitation downscaling task.  ... 
arXiv:2012.09700v2 fatcat:7c3vhna7rjfgpcitfilomhiwki

Evaluating the Clinical Realism of Synthetic Chest X-Rays Generated Using Progressively Growing GANs

Bradley Segal, David M. Rubin, Grace Rubin, Adam Pantanowitz
2021 SN Computer Science  
We release our source code, model weights, and an archive of labelled generates.  ...  We confirm these findings by evaluating synthetic classification model performance on real scans.  ...  Human eYe Perceptual Evaluation (HYPE) [56] in contrast to the aforementioned automated approaches, standardises the human evaluation of model generates by considering either the time necessary to discriminate  ... 
doi:10.1007/s42979-021-00720-7 pmid:34104898 pmcid:PMC8176276 fatcat:p2tgy6s3y5dsfoyebe5saoo3my

Evaluating the Clinical Realism of Synthetic Chest X-Rays Generated Using Progressively Growing GANs [article]

Bradley Segal, David M. Rubin, Grace Rubin, Adam Pantanowitz
2021 arXiv   pre-print
We apply a PGGAN to the task of unsupervised x-ray synthesis and have radiologists evaluate the clinical realism of the resultant samples.  ...  We confirm these findings by evaluating synthetic classification model performance on real scans.  ...  Human eYe Perceptual Evaluation (HYPE) [45] in contrast to the aforementioned automated approaches, standardises the human evaluation of model generates by considering either the time necessary to discriminate  ... 
arXiv:2010.03975v2 fatcat:7incgpatqzfjri2jvyljadv5bm

Deep Nets: What have they ever done for Vision? [article]

Alan L. Yuille, Chenxi Liu
2020 arXiv   pre-print
We argue that this combinatorial explosion takes us into a regime where "big data is not enough" and where we need to rethink our methods for benchmarking performance and evaluating vision algorithms.  ...  At present Deep Nets perform very well on specific visual tasks with benchmark datasets but they are much less general purpose, flexible, and adaptive than the human visual system.  ...  of these topics.  ... 
arXiv:1805.04025v4 fatcat:dtaztvqtmjbdrmzxkat36ear2m

A State-of-the-Art Review on Image Synthesis with Generative Adversarial Networks

Lei Wang, Wei Chen, Wenjia Yang, Fangming Bi, Fei Richard Yu
2020 IEEE Access  
Generative Adversarial Networks (GANs) have achieved impressive results in various image synthesis tasks, and are becoming a hot topic in computer vision research because of the impressive performance  ...  We also preview some likely future research directions in the field of GANs, such as video generation, facial animation synthesis and 3D face reconstruction.  ...  [111] proposed a method called HUMAN EYE PERCEPTUAL EVALUATION (HYPE) to establish a gold standard human benchmark for generative realism. Grnarova et al.  ... 
doi:10.1109/access.2020.2982224 fatcat:p5uxjh4cybfw5grp6ldhkpukrm

Assessing the impact of machine intelligence on human behaviour: an interdisciplinary endeavour [article]

Emilia Gómez, Carlos Castillo, Vicky Charisi, Verónica Dahl, Gustavo Deco, Blagoj Delipetrev, Nicole Dewandre, Miguel Ángel González-Ballester, Fabien Gouyon, José Hernández-Orallo, Perfecto Herrera, Anders Jonsson, Ansgar Koene (+11 others)
2018 arXiv   pre-print
The workshop was organized in the context of a new research programme at the Centre for Advanced Studies, Joint Research Centre of the European Commission, which focuses on studying the potential impact  ...  intelligence, algorithm's potential impact on human cognitive capabilities and decision making, and evaluation and regulation needs.  ...  Acknowledgements This workshop was organized by the Centre for Advanced Studies, Joint Research Centre with the local support of the Department of Information and Communication Technologies, Universitat  ... 
arXiv:1806.03192v1 fatcat:jypjtwtgxbg5vbkigudnxtsp64

Monocular Quasi-Dense 3D Object Tracking [article]

Hou-Ning Hu, Yung-Hsu Yang, Tobias Fischer, Trevor Darrell, Fisher Yu, Min Sun
2021 arXiv   pre-print
A reliable and accurate 3D tracking framework is essential for predicting future locations of surrounding objects and planning the observer's actions in numerous applications such as autonomous driving  ...  We propose a framework that can effectively associate moving objects over time and estimate their full 3D bounding box information from a sequence of 2D images captured on a moving platform.  ...  Thanks to the hype of autonomous driving, more and more large-scale datasets, and benchmarks have surged onto the desk.  ... 
arXiv:2103.07351v1 fatcat:tn7il5h4gbcrxdl5kvb7otgzcu

Phronesis of AI in radiology: Superhuman meets natural stupidity [article]

Judy W. Gichoya, Siddhartha Nuthakki, Pallavi G. Maity, Saptarshi Purkayastha
2018 arXiv   pre-print
We performed one of the first studies in radiology to see how human and AI performance can complement and improve each others performance for detecting pneumonia in chest X-rays.  ...  We review 'superhuman' AI performance claims in radiology and then provide a self-reflection on our own work in the area in the form of a critical review, a tribute of sorts to McDermotts 1976 paper, asking  ...  Acknowledgments Thanks to the radiology residents and fellows' journal club participants for being willing participants to participate in the human versus machine competition.  ... 
arXiv:1803.11244v1 fatcat:tblgayob5jegzjzfauhgrtx6gq

Using AI Methods to Evaluate a Minimal Model for Perception

Robert Prentner, Chris Fields
2019 Open Philosophy  
We conjecture that a combination of AI research and philosophy allows general principles of mind and being to emerge from a "quasi-empirical" investigation.  ...  perceptual inference?  ...  We thank our colleagues for fruitful discussions that were related to this paper. In particular, we thank Chetan Prakash for helpful comments on an earlier version.  ... 
doi:10.1515/opphil-2019-0034 fatcat:4olubc54a5crnj72bpcrzcyh6i

Selected Tracking Technologies and Sensor Integrations witin the Easy Reading framework

Stefan Parker
2019 Zenodo  
This deliverable gives a short overview of the state of the art in user tracking by discussing a few examples of existing systems for web tracking (Google Analytics) and emotion detection (Face Reader,  ...  It then gives an overview of existing related EU funded research, also giving an "Of interest rating" from the view of the Easy Reading project.  ...  The proposed project aims to fill this gap in our understanding of human language, by combining computational modelling and human experimental research. [---] Next, in eye-tracking and pupillometry studies  ... 
doi:10.5281/zenodo.3375228 fatcat:wsrktm4rwfcrhouvx25kmrmz2m

Neural Networks and Neuroscience-Inspired Computer Vision

David Daniel Cox, Thomas Dean
2014 Current Biology  
They transform a torrent of complex and ambiguous sensory information into coherent thought and action, allowing an organism to perceive and model its environment, synthesize and make decisions from disparate  ...  streams of information, and adapt to a changing environment.  ...  A related issue for modern computer vision is out-of-set generalization. One fundamental challenge in computer vision is in evaluating performance.  ... 
doi:10.1016/j.cub.2014.08.026 pmid:25247371 fatcat:wjw5ewsfsfbprgx42yudsrxsse

Backwards is the way forward: Feedback in the cortical hierarchy predicts the expected future

Lars Muckli, Lucy S. Petro, Fraser W. Smith
2013 Behavioral and Brain Sciences  
Second, hierarchical prediction offers progress on aconcretedescriptive level for testing and constraining conceptual elements and mechanisms of predictive coding models (estimation of predictions, prediction  ...  AbstractClark offers a powerful description of the brain as a prediction machine, which offers progress on two distinct levels.  ...  But at this stage, there is much to fill in for it to function as a general model of human cognition and action.  ... 
doi:10.1017/s0140525x12002361 pmid:23663531 fatcat:d7iicta6qfau5cb3znkiwtk4g4

Whatever next? Predictive brains, situated agents, and the future of cognitive science

2013 Behavioral and Brain Sciences  
This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing.  ...  Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to adaptive success.  ...  But at this stage, there is much to fill in for it to function as a general model of human cognition and action.  ... 
doi:10.1017/s0140525x12000477 pmid:23663408 fatcat:k7kqhjt5vngdfjftnixoef5bxq
« Previous Showing results 1 — 15 out of 192 results