ABSTRACT
Refractive errors, such as myopia and astigmatism, can lead to severe visual impairment if not detected and corrected in time. Traditional methods of refractive error diagnosis rely on well-trained optometrists operating expensive and importable devices, constraining the vision screening process. Advance in smartphone camera has enabled novel low-cost ubiquitous vision screening to detect refractive error or ametropia through eye image processing, based on the principle of photorefraction. However, contemporary smartphone-based methods rely heavily on hand-crafted features and sufficiency of well-labeled data. To address these challenges, this paper exploits active learning methods with a set of Convolutional Neural Network features encoding information of human eyes from pre-trained gaze estimation model. This enables more effective training on refractive error detection models with less labeled data. Our experimental results demonstrate the encouraging effectiveness of our active learning approach. The new set of features is able to attain screening accuracy of more than 80% with mean absolute error less than 0.66, meeting the expectation of optometrists for 0.5 to 1. The proposed active learning also requires significantly fewer training samples of 18% in achieving satisfactory performance.
Supplemental Material
- Robert W Arnold, James W O'Neil, Kim L Cooper, David I Silbert, and Sean P Donahue. 2018. Evaluation of a smartphone photoscreening app to detect refractive amblyopia risk factors in children aged 1--6 years. Clinical Ophthalmology (Auckland, NZ), Vol. 12 (2018), 1533.Google Scholar
- Ujjwal Baid, Bhakti Baheti, Prasad Dutande, and Sanjay Talbar. 2019. Detection of Pathological Myopia and Optic Disc Segmentation with Deep Convolutional Neural Networks. In TENCON 2019--2019 IEEE Region 10 Conference (TENCON). IEEE, 1345--1350.Google Scholar
- William H Beluch, Tim Genewein, Andreas Nürnberger, and Jan M Köhler. 2018. The power of ensembles for active learning in image classification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 9368--9377.Google ScholarCross Ref
- WR Bobier and OJ Braddick. 1985. Eccentric photorefraction: optical analysis and empirical measures. American journal of optometry and physiological optics, Vol. 62, 9 (1985), 614--620.Google Scholar
- Leo Breiman. 1996. Bagging predictors. Machine learning, Vol. 24, 2 (1996), 123--140.Google Scholar
- Gulcan Can, Yassir Benkhedda, and Daniel Gatica-Perez. 2018. Ambiance in social media venues: visual cue interpretation by machines and crowds. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. 2363--2372.Google ScholarCross Ref
- Osbert YC Chan, Marion Edwards, and Brian Brown. 1996. Calibration and validity of an eccentric photorefractor. Ophthalmic and Physiological Optics, Vol. 16, 3 (1996), 203--210.Google ScholarCross Ref
- Chih-Chung Chang and Chih-Jen Lin. 2011. LIBSVM: A library for support vector machines. ACM transactions on intelligent systems and technology (TIST), Vol. 2, 3 (2011), 1--27.Google ScholarDigital Library
- Fouzi Douak, Farid Melgani, Edoardo Pasolli, and Nabil Benoudjit. 2012. SVR active learning for product quality control. In 2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA). IEEE, 1113--1117.Google ScholarCross Ref
- Yoav Freund and Robert E Schapire. 1995. A desicion-theoretic generalization of on-line learning and an application to boosting. In European conference on computational learning theory. Springer, 23--37.Google ScholarDigital Library
- Yarin Gal, Riashat Islam, and Zoubin Ghahramani. 2017. Deep bayesian active learning with image data. In Proceedings of the 34th International Conference on Machine Learning-Volume 70. JMLR. org, 1183--1192.Google ScholarDigital Library
- Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531 (2015).Google Scholar
- Howard C Howland. 1985. Optics of photoretinoscopy: results from ray tracing. American journal of optometry and physiological optics, Vol. 62, 9 (1985), 621--625.Google Scholar
- OA Hunt, JS Wolffsohn, and B Gilmartin. 2003. Evaluation of the measurement of refractive error by the PowerRefractor: a remote, continuous and binocular measurement system of oculomotor function. British Journal of Ophthalmology, Vol. 87, 12 (2003), 1504--1508.Google ScholarCross Ref
- Rebecca Hwa. 2004. Sample selection for statistical parsing. Computational linguistics, Vol. 30, 3 (2004), 253--276.Google Scholar
- Kari Kaakinen and VEIKKO TOMMILA. 1979. A clinical study on the detection of strabismus, anisometropia or ametropia of children by simultaneous photography of the corneal and the fundus reflexes. Acta ophthalmologica, Vol. 57, 4 (1979), 600--611.Google Scholar
- Kyle Krafka, Aditya Khosla, Petr Kellnhofer, Harini Kannan, Suchendra Bhandarkar, Wojciech Matusik, and Antonio Torralba. 2016. Eye tracking for everyone. In Proceedings of the IEEE conference on computer vision and pattern recognition. 2176--2184.Google ScholarCross Ref
- Tiffany CK Kwok, Naomi CM Shum, Grace Ngai, Hong Va Leong, Grace Amy Tseng, Hoi-yi Choi, Ka-yan Mak, and Chi-Wai Do. 2015. Democratizing Optometric Care: A Vision-Based, Data-Driven Approach to Automatic Refractive Error Measurement for Vision Screening. In 2015 IEEE International Symposium on Multimedia (ISM). IEEE, 7--12.Google ScholarCross Ref
- Chen Liang, Jianbo Ye, Shuting Wang, Bart Pursel, and C Lee Giles. 2018. Investigating active learning for concept prerequisite learning. In Thirty-Second AAAI Conference on Artificial Intelligence.Google Scholar
- Naoki Abe Hiroshi Mamitsuka et almbox. 1998. Query learning strategies using boosting and bagging. In Machine learning: proceedings of the fifteenth international conference (ICML'98), Vol. 1. Morgan Kaufmann Pub.Google Scholar
- Prem Melville and Raymond J Mooney. 2004. Diverse ensembles for active learning. In Proceedings of the twenty-first international conference on Machine learning. 74.Google ScholarDigital Library
- AC Molteno, J Hoare-Nairne, JC Parr, Anne Simpson, IJ Hodgkinson, NE O'Brien, and SD Watts. 1983. The Otago photoscreener, a method for the mass screening of infants to detect squint and refractive errors. Transactions of the Ophthalmological Society of New Zealand, Vol. 35 (1983), 43--49.Google Scholar
- Seonwook Park, Adrian Spurr, and Otmar Hilliges. 2018a. Deep pictorial gaze estimation. In Proceedings of the European Conference on Computer Vision (ECCV). 721--738.Google ScholarDigital Library
- Seonwook Park, Xucong Zhang, Andreas Bulling, and Otmar Hilliges. 2018b. Learning to find eye region landmarks for remote gaze estimation in unconstrained settings. In Proceedings of the 2018 ACM Symposium on Eye Tracking Research & Applications. 1--10.Google ScholarDigital Library
- Donatella Pascolini and Silvio Paolo Mariotti. 2012. Global estimates of visual impairment: 2010. British Journal of Ophthalmology, Vol. 96, 5 (2012), 614--618.Google ScholarCross Ref
- M M. W Peterseim, R. S Rhodes, R. N Patel, M E. Wilson, L. E Edmondson, S. A Logan, E. W Cheeseman, E. Shortridge, and R. H Trivedi. 2018. Effectiveness of the GoCheck Kids vision screener in detecting amblyopia risk factors. American journal of ophthalmology, Vol. 187 (2018), 87--91.Google Scholar
- Michael X Repka, Raymond T Kraker, Jonathan M Holmes, Allison I Summers, Stephen R Glaser, Carmen N Barnhardt, and David R Tien. 2014. Atropine vs patching for treatment of moderate amblyopia: follow-up at 15 years of age of a randomized clinical trial. JAMA ophthalmology, Vol. 132, 7 (2014), 799--805.Google Scholar
- Clara I Sánchez, Meindert Niemeijer, Michael D Abràmoff, and Bram van Ginneken. 2010. Active learning for an efficient training strategy of computer-aided diagnosis systems: application to diabetic retinopathy screening. In International Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 603--610.Google ScholarCross Ref
- Frank Schaeffel, Howard C Howland, and Leslie Farkas. 1986. Natural accommodation in the growing chicken. Vision Research, Vol. 26, 12 (1986), 1977--1993.Google ScholarCross Ref
- Ramprasaath R Selvaraju, Abhishek Das, Ramakrishna Vedantam, Michael Cogswell, Devi Parikh, and Dhruv Batra. 2016. Grad-CAM: Why did you say that? arXiv preprint arXiv:1611.07450 (2016).Google Scholar
- Burr Settles. 2009. Active learning literature survey. Technical Report. University of Wisconsin-Madison Department of Computer Sciences.Google ScholarDigital Library
- Burr Settles and Mark Craven. 2008. An analysis of active learning strategies for sequence labeling tasks. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing. 1070--1079.Google ScholarCross Ref
- Amy L Sheppard and Leon N Davies. 2010. Clinical evaluation of the Grand Seiko auto ref/keratometer WAM-5500. Ophthalmic and Physiological Optics, Vol. 30, 2 (2010), 143--151.Google ScholarCross Ref
- Jennifer Vandoni, Emanuel Aldea, and Sylvie Le Hégarat-Mascle. 2019. Evidential query-by-committee active learning for pedestrian detection in high-density crowds. International Journal of Approximate Reasoning, Vol. 104 (2019), 166--184.Google ScholarCross Ref
- Avinash V Varadarajan, Ryan Poplin, Katy Blumer, Christof Angermueller, Joe Ledsam, Reena Chopra, Pearse A Keane, Greg S Corrado, Lily Peng, and Dale R Webster. 2018. Deep learning for predicting refractive error from retinal fundus images. Investigative ophthalmology & visual science, Vol. 59, 7 (2018), 2861--2868.Google Scholar
- Xuehan Xiong and Fernando De la Torre. 2013. Supervised descent method and its applications to face alignment. In Proceedings of the IEEE conference on computer vision and pattern recognition. 532--539.Google ScholarDigital Library
- Xucong Zhang, Yusuke Sugano, Mario Fritz, and Andreas Bulling. 2017. Mpiigaze: Real-world dataset and deep appearance-based gaze estimation. IEEE transactions on pattern analysis and machine intelligence, Vol. 41, 1 (2017), 162--175.Google Scholar
- Bolei Zhou, Aditya Khosla, Agata Lapedriza, Aude Oliva, and Antonio Torralba. 2016. Learning deep features for discriminative localization. In Proceedings of the IEEE conference on computer vision and pattern recognition. 2921--2929.Google ScholarCross Ref
Index Terms
- Exploiting Active Learning in Novel Refractive Error Detection with Smartphones
Recommendations
Screening for refractive error with low-quality smartphone images
MoMM '20: Proceedings of the 18th International Conference on Advances in Mobile Computing & MultimediaUncorrected refractive errors can lead to permanent debilitating eye conditions if not corrected in a timely manner. Contemporary diagnostic methods rely on the professional acumen of optometrists and the use of expensive devices, which may not be ...
Hand-eye Coordination for Textual Difficulty Detection in Text Summarization
ICMI '20: Proceedings of the 2020 International Conference on Multimodal InteractionThe task of summarizing a document is a complex task that requires a person to multitask between reading and writing processes. Since a person's cognitive load during reading or writing is known to be dependent upon the level of comprehension or ...
NETRA: interactive display for estimating refractive errors and focal range
SIGGRAPH '10: ACM SIGGRAPH 2010 papersWe introduce an interactive, portable, and inexpensive solution for estimating refractive errors in the human eye. While expensive optical devices for automatic estimation of refractive correction exist, our goal is to greatly simplify the mechanism by ...
Comments