Abstract
Large amounts of suitably marked images are required to feed a learning algorithm when building a detector. The process of collecting images and then marking locations of target objects in them is arduous. One could potentially speed up this process if real images for a given application were replaced by images generated synthetically, and coordinates of targets were simply imposed, rather then discovered manually. Despite the appeal of such an automatization, questions arise regarding the usefulness of systems built this way in real operating conditions. In particular, the obvious violation of i.i.d. principle might result in higher error rates at the testing stage.
In this paper we provide an experimental study of the above approach, taking up road sign detection as an example. We generate synthetic training scenes by laying road sign icons randomly over a set of backgrounds with additional perturbations (rotations, brightness changes, blurring, sharpening, noise). Ensemble learning is then carried out using a RealBoost algorithm with shallow decision trees. Haar-like features or Fourier moments constitute the direct input information extracted from images. In both cases we support the computations with suitable integral images. Accuracy of resulting detectors is finally tested on real images.
This work was financed by the National Science Centre, Poland. Research project no.: 2016/21/B/ST6/01495.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
As regards the negative examples, they usually do not have to be processed manually. One can implement an automatic sampling procedure that picks negative windows on random e.g. from images containing no targets.
- 2.
Probably Approximately Correct.
- 3.
An input image has to be scanned with a sliding window at several scales. Commonly, more than \(10^4\) window positions are checked for typical settings.
- 4.
Templates can be scaled (stretched or squeezed) along either axis.
- 5.
\(\langle a, b \rangle =\mathop {\sum \sum }\nolimits _{x,y} a(x,y) \overline{b}(x,y)\), where \(\overline{b}\) denotes the complex conjugate of b.
- 6.
The moment of order zero is a real number: .
References
Anthony, M., Bartlett, P.: Neural Network Learning: Theoretical Foundations. Cambridge University Press, Cambridge (2009)
Bera, A., Klęsk, P., Sychel, D.: Constant-time calculation of zernike moments for detection with rotational invariance. IEEE Trans. Pattern Anal. Mach. Intell. 41(3), 537–551 (2019)
Cherkassky, V., Mulier, F.: Learning from Data, 2nd edn. Wiley, Hoboken (2007)
Friedman, J., Hastie, T., Tibshirani, R.: Additive logistic regression: a statistical view of boosting. Ann. Stat. 28(2), 337–407 (2000)
Klęsk, P., Kapruziak, M., Olech, B.: Fast extraction of 3D fourier moments via multiple integral images: an application to antitank mine detection in GPR C-Scans. In: Chmielewski, L.J., Datta, A., Kozera, R., Wojciechowski, K. (eds.) ICCVG 2016. LNCS, vol. 9972, pp. 206–220. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-46418-3_19
Klȩsk, P.: Constant-time fourier moments for face detection—can accuracy of haar-like features be beaten? In: Rutkowski, L., Korytkowski, M., Scherer, R., Tadeusiewicz, R., Zadeh, L.A., Zurada, J.M. (eds.) ICAISC 2017. LNCS (LNAI), vol. 10245, pp. 530–543. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-59063-9_47
Klęsk, P., Godziuk, A., Kapruziak, M., Olech, B.: Fast analysis of C-Scans from ground penetrating radar via 3D Haar-like features with application to landmine detection. IEEE Trans. Geosci. Remote Sens. 53(7), 3996–4009 (2015)
Korytkowski, M., Rutkowski, L., Scherer, R.: Fast image classification by boosting fuzzy classifiers. Inf. Sci. 327, 175–182 (2016)
Montserrat, D., et al.: Logo detection and recognition with synthetic images. Electronic Imaging 7, 337-1–337-7 (2018). Imaging and Multimedia Analytics in a Web and Mobile World 2018
Rozantsev, A., Lepetit, V., Fua, P.: On rendering synthetic images for training an object detector. Comput. Vis. Image Underst. 137, 24–37 (2015)
Vapnik, V.: Statistical Learning Theory: Inference from Small Samples. Wiley, New York (1998)
Viola, P., Jones, M.: Rapid object detection using a boosted cascade of simple features. In: Conference on Computer Vision and Pattern Recognition, CVPR 2001, pp. 511–518. IEEE (2001)
Zhang, Q., et al.: Wildland forest fire smoke detection based on faster R-CNN using synthetic smoke images. Procedia Eng. 211, 441–446 (2018). 2017 8th International Conference on Fire Science and Fire Protection Engineering (ICFSFPE 2017)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Klęsk, P. (2019). From Synthetic Images Towards Detectors of Real Objects: A Case Study on Road Sign Detection. In: Rutkowski, L., Scherer, R., Korytkowski, M., Pedrycz, W., Tadeusiewicz, R., Zurada, J. (eds) Artificial Intelligence and Soft Computing. ICAISC 2019. Lecture Notes in Computer Science(), vol 11509. Springer, Cham. https://doi.org/10.1007/978-3-030-20915-5_2
Download citation
DOI: https://doi.org/10.1007/978-3-030-20915-5_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-20914-8
Online ISBN: 978-3-030-20915-5
eBook Packages: Computer ScienceComputer Science (R0)