Cognitive Relevance Transform for Population Re-Targeting
Abstract
:1. Introduction
Scope of the Research
2. Related Work
3. The Cognitive Relevance Transform
Algorithm 1 The Cognitive Relevance Transform. |
The Cognitive Relevance Transform (CRT) is a sequence of operations in image categories. The CRT is obtained in three steps. First (operation 1), the dataset is subsampled by Algorithm 2. Second (operation 2), user studies are performed (see Section 3.2 and Algorithm 3). Third (operation 3), merging operations E and separation operations S are determined by Algorithm 4. Renaming operations R are determined by Algorithm 5. = denotes definition and ← denotes calculation. |
Require: |
Ensure: CRT |
1: dataset subsampling(D) |
2: user studies() |
3: get crt() |
4: function get crt() |
5: determining merging and separation operations() |
6: determining renaming operations(A) |
7: return |
8: end function |
3.1. Reduction of the Dataset Size
Algorithm 2 Dataset Subsampling. |
To reduce the dataset size, the minimum number of categories is calculated by the statistical power of two-sample test for proportions with unequal sample sizes. For details, see [35]. The minimum number of images per category is calculated by the statistical power of statistical test of independence [35]. Both, and are then corrected to a discrete number by operation 6. Finally, the number of batches is calculated. ← denotes calculation. |
Require: |
Ensure: |
1:pwr.2p2n.test() |
2: pwr.chisq.test() |
3: to integer() |
4: to integer() |
5: |
6: functionto integer() |
7: return |
8: end function |
3.2. User Studies
Algorithm 3 Population size. |
Effect size for the number of observations is calculated by operation 1 (see [35]). Next, is determined by the statistical power of the binomial test [35]. Finally, the number of batches and the number of human subjects are calculated. ← denotes calculation. |
Require: |
Ensure: |
1: |
2: pwr.p.test() |
3: |
3.3. Deriving CRT operations
Algorithm 4 Merging and Separation operations. |
To determine merging operations E and separation operations S, a set of feature vectors X is clustered into a set of C clusters. Then, for each cluster and category get as a relative frequency of clustered into . Use to separate the category . In context of a cluster , separation of is denoted by . A set of separation operations for all categories in the context of a cluster is denoted by . Merging operation in the context of cluster is calculated by merging results from separation operations . = denotes definition and ← denotes calculation. |
Require: |
Ensure: |
1: clustering(X) |
2: get crt operations() |
3: function get crt operations() |
4: for do |
5: for do |
6: frequency() |
7: separate() |
8: end for |
9: |
10: merge() |
11: end for |
12: return |
13: end function |
Algorithm 5 Determining Renaming operations. |
A set of labels A is cleaned by the number of text operations. Cleaned labels are structured into a set of feature vectors Y and clustered into a set of Z clusters. Renaming operation for cluster is determined by the most frequent label. = denotes definition and ← denotes calculation. |
Require: |
Ensure:R |
1: clean(A) |
2: |
3: clustering(Y) |
4: most frequent label(Z) |
5: function clean(A) |
6: to lowercase(A) |
7: decode to closest ASCII code() |
8: remove non-alphabetic characters() |
9: abbreviations to whole words() |
10: strip white space() |
11: word segmentation() |
12: tokenize words() |
13: keep nouns, adjectives and prepositions() |
14: spell check() |
15: lemmatize() |
16: return |
17: end function |
4. User Population Re-Targeting
Algorithm 6 Human subject consensus. |
Individual answers by human subjects L are transformed to the population answers . First, calculate number of successes and trials. Success is a human label that is equivalent to original label . A binomial test (for details see [35]) is used to determine consensus label as a population answer. If more than human subject consensus exists, consensus label becomes the most frequent label , otherwise it becomes the original label . = denotes definition, ≡ denotes equivalence, and ← denotes calculation. |
Require: |
Ensure: |
1: for do |
2: successes |
3: trials |
4: the most frequent label() |
5: p-value ←binomial test(successes, trials, ) |
6: if p-value < then |
7: |
8: else |
9: |
10: end if |
11: end for |
5. Experiments
5.1. Materials
5.2. Experiments on ILSVRC2012 Dataset
5.2.1. Reduction of Dataset Size
5.2.2. User Studies
5.2.3. Choice Of Categories
5.2.4. Deep Learning Models
5.3. Experiments on VireoFood-172 Dataset
5.3.1. Reduction of the Dataset Size
5.3.2. User Studies
5.3.3. Choice of Categories
5.3.4. Deep Learning Models
6. ILSVRC2012 Results and Discussion
6.1. Deriving CRT operations
6.2. Human Population and Machine Classification after Applying the CRT
6.3. Qualitative Results
6.4. Verification
7. VireoFood-172 Results and Discussion
7.1. Training CNNs
7.2. Pre-CRT Results
7.3. Deriving CRT Operations
7.4. Post-CRT Results
7.5. Qualitative Results
8. Conclusions
Summary of Findings
- •
- The performance of human observers on ImageNet problem is probably better than previously thought.
- •
- The performance of CNNs is target user population-dependent.
- •
- Not all dataset biases are created equal.
- •
- User testing in the evaluation of CNNs is needed.
- •
- In many cases, mistakes, made by CNNs are grave.
Author Contributions
Funding
Conflicts of Interest
Dataset
Abbreviations
AI | artificial intelligence |
CNN | Convolutional Neural Network |
HMO | hierarchical neural network model |
ID | identification number |
ILSVRC2012 | ImageNet Large Scale Visual Recognition Challenge 2012 |
mAP | mean average precision |
ACC | accuracy |
CFM | confusion matrix |
CRT | cognitive relevance transform |
pp | percentage points |
ENG | English population |
ASIA | Asian population |
Appendix A. ILSVRC2012 Results
Appendix B. VireoFood-172 Results
References
- Guo, Y.; Liu, Y.; Oerlemans, A.; Lao, S.; Wu, S.; Lew, M.S. Deep learning for visual understanding: A review. Neurocomputing 2016, 187, 27–48. [Google Scholar] [CrossRef]
- Fleuret, F.; Li, T.; Dubout, C.; Wampler, E.K.; Yantis, S.; Geman, D. Comparing machines and humans on a visual categorization test. Proc. Nat. Acad. Sci. USA 2011, 108, 17621–17625. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Lake, B.M.; Salakhutdinov, R.; Tenenbaum, J.B. Human-level concept learning through probabilistic program induction. Science 2015, 350, 1332–1338. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Pramod, R.T.; Arun, S.P. Do Computational Models Differ Systematically from Human Object Perception? In Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; pp. 1601–1609. [Google Scholar]
- Geirhos, R.; Janssen, D.H.J.; Schütt, H.H.; Rauber, J.; Bethge, M.; Wichmann, F.A. Comparing deep neural networks against humans: Object recognition when the signal gets weaker. arXiv 2017, arXiv:1706.06969. [Google Scholar]
- Lake, B.M.; Ullman, T.D.; Tenenbaum, J.B.; Gershman, S.J. Building Machines That Learn and Think Like People. Behav. Brain Sci. 2017, 40, e253. [Google Scholar] [CrossRef] [Green Version]
- Das, A.; Agrawal, H.; Zitnick, L.; Parikh, D.; Batra, D. Human Attention in Visual Question Answering: Do Humans and Deep Networks Look at the Same Regions? Comput. Vis. Image Underst. 2017, 163, 90–100. [Google Scholar] [CrossRef] [Green Version]
- Linsley, D.; Eberhardt, S.; Sharma, T.; Gupta, P.; Serre, T. What are the visual features underlying human versus machine vision? In Proceedings of the 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), Venice, Italy, 22–29 October 2017; pp. 2706–2714. [Google Scholar]
- Norman, D. The Design of Everyday Things: Revised and Expanded Edition; Basic Books: New York, NY, USA, 2013. [Google Scholar]
- Geirhos, R.; Rubisch, P.; Michaelis, C.; Bethge, M.; Wichmann, F.A.; Brendel, W. ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. In Proceedings of the Seventh International Conference on Learning Representations (ICLR), New Orleans, LA, USA, 6–9 May 2019. [Google Scholar]
- Deng, J.; Dong, W.; Socher, R.; Li, L.J.; Li, K.; Fei-Fei, L. ImageNet: A large-scale hierarchical image database. In Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Miami, FL, USA, 20–25 June 2009; pp. 248–255. [Google Scholar]
- Wilson, D.; Sperber, D. Relevance Theory. In The Handbook of Pragmatics; Blackwell: Oxford, UK, 2004; Chapter 27; pp. 607–632. [Google Scholar]
- Torralba, A.; Efros, A.A. Unbiased Look at Dataset Bias. In Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Colorado Springs, CO, USA, 20–25 June 2011; pp. 1521–1528. [Google Scholar]
- Tommasi, T.; Patricia, N.; Caputo, B.; Tuytelaars, T. A Deeper Look at Dataset Bias. In Domain Adaptation in Computer Vision Applications; Springer: Berlin/Heidelberg, Germany, 2017; pp. 37–55. [Google Scholar]
- Zendel, O.; Honauer, K.; Murschitz, M.; Humenberger, M.; Domínguez, G.F. Analyzing Computer Vision Data—The Good, the Bad and the Ugly. In Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, 21–26 July 2017; pp. 6670–6680. [Google Scholar]
- Ponce, J.; Berg, T.L.; Everingham, M.; Forsyth, D.A.; Hebert, M.; Lazebnik, S.; Marszalek, M.; Schmid, C.; Russell, B.C.; Torralba, A.; et al. Dataset Issues in Object Recognition. In Toward Category-Level Object Recognition; Springer: Berlin/Heidelberg, Germany, 2006; pp. 29–48. [Google Scholar]
- Pinto, N.; Cox, D.D.; DiCarlo, J.J. Why is Real-World Visual Object Recognition Hard? PLoS Comput. Biol. 2008, 4, e27. [Google Scholar] [CrossRef]
- Lakoff, G. Women, Fire, and Dangerous Things; The University of Chicago Press: Chicago, IL, USA, 1987; p. 632. [Google Scholar]
- Russakovsky, O.; Deng, J.; Su, H.; Krause, J.; Satheesh, S.; Ma, S.; Huang, Z.; Karpathy, A.; Khosla, A.; Bernstein, M.; et al. ImageNet Large Scale Visual Recognition Challenge. Int. J. Comput. Vis. 2015, 115, 211–252. [Google Scholar] [CrossRef] [Green Version]
- Stock, P.; Cisse, M. ConvNets and ImageNet Beyond Accuracy: Understanding Mistakes and Uncovering Biases. In Computer Vision—ECCV 2018; Springer: Munich, Germany, 2018; pp. 504–519. [Google Scholar]
- Marcus, G. Deep Learning: A Critical Appraisal. arXiv 2018, arXiv:cs.AI/1801.00631. [Google Scholar]
- Nguyen, A.; Yosinski, J.; Clune, J. Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7–12 June 2015; pp. 427–436. [Google Scholar]
- Szegedy, C.; Zaremba, W.; Sutskever, I.; Bruna, J.; Erhan, D.; Goodfellow, I.; Fergus, R. Intriguing properties of neural networks. In Proceedings of the 2nd International Conference on Learning Representations (ICLR), Banff, AB, Canada, 14–16 April 2014. [Google Scholar]
- Goodfellow, I.J.; Shlens, J.; Szegedy, C. Explaining and Harnessing Adversarial Examples. In Proceedings of the 3rd International Conference on Learning Representations (ICLR), San Diego, CA, USA, 7–9 May 2015. [Google Scholar]
- Tabacof, P.; Valle, E. Exploring the Space of Adversarial Images. In Proceedings of the 2016 International Joint Conference on Neural Networks (IJCNN), Vancouver, BC, Canada, 24–29 July 2016; pp. 426–433. [Google Scholar]
- Dodge, S.; Karam, L. A Study and Comparison of Human and Deep Learning Recognition Performance Under Visual Distortions. In Proceedings of the 2017 26th International Conference on Computer Communication and Networks (ICCCN), Vancouver, BC, Canada, 31 July–3 August 2017; pp. 1–7. [Google Scholar]
- Dodge, S.; Karam, L. Can the Early Human Visual System Compete with Deep Neural Networks? In Proceedings of the 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), Venice, Italy, 22–29 October 2017; pp. 2798–2804. [Google Scholar]
- Wichmann, F.A.; Janssen, D.H.J.; Geirhos, R.; Aguilar, G.; Schütt, H.H.; Maertens, M.; Bethge, M. Methods and measurements to compare men against machines. Electron. Imaging 2017, 14, 36–45. [Google Scholar] [CrossRef]
- Kheradpisheh, S.R.; Ghodrati, M.; Ganjtabesh, M.; Masquelier, T. Deep Networks Can Resemble Human Feed-forward Vision in Invariant Object Recognition. Sci. Rep. 2016, 6, 32672. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Kheradpisheh, S.R.; Ghodrati, M.; Ganjtabesh, M.; Masquelier, T. Humans and Deep Networks Largely Agree on Which Kinds of Variation Make Object Recognition Harder. Front. Comput. Neurosci. 2016, 10, 92. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Stabinger, S.; Rodríguez-Sánchez, A.; Piater, J. 25 years of CNNs: Can we compare to human abstraction capabilities? In International Conference on Artificial Neural Networks; Springer International Publishing: Cham, Switzerland, 2016; pp. 380–387. [Google Scholar]
- Yamins, D.L.K.; Hong, H.; Cadieu, C.F.; Solomon, E.A.; Seibert, D.; DiCarlo, J.J. Performance-optimized hierarchical models predict neural responses in higher visual cortex. Proc. Nat. Acad. Sci. USA 2014, 111, 8619–8624. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Cadieu, C.F.; Hong, H.; Yamins, D.L.K.; Pinto, N.; Ardila, D.; Solomon, E.A.; Majaj, N.J.; DiCarlo, J.J. Deep Neural Networks Rival the Representation of Primate IT Cortex for Core Visual Object Recognition. PLoS Comput. Biol. 2014, 10, e1003963. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Rajalingham, R.; Issa, E.B.; Bashivan, P.; Kar, K.; Schmidt, K.; DiCarlo, J.J. Large-scale, high-resolution comparison of the core visual object recognition behavior of humans, monkeys, and state-of-the-art deep artificial neural networks. J. Neurosci. 2018, 38, 7255–7269. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Lawrence Erlbaum Associates: Hillsdale, NJ, USA, 1988; p. 567. [Google Scholar]
- Allaire, J. RStudio: Integrated Development Environment for R; RStudio, Inc.: Boston, MA, USA, 2016. [Google Scholar]
- Team, R.C. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2016. [Google Scholar]
- Champely, S.; Ekstrom, C.; Dalgaard, P.; Gill, J.; Weibelzahl, S.; De Rosario, H. pwr: Basic Functions for Power Analysis; R Foundation for Statistical Computing: Vienna, Austria, 2016. [Google Scholar]
- Faul, F.; Erdfelder, E.; Lang, A.G.; Buchner, A. G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods 2007, 39, 175–191. [Google Scholar] [CrossRef]
- De Vos, N. Kmodes. 2016. Available online: https://github.com/nicodv/kmodes (accessed on 21 May 2020).
- Huang, Z. Clustering large data sets with mixed numeric and categorical values. In Proceedings of the First Pacific Asia Knowledge Discovery and Data Mining Conference (PAKDD), Singapore, 23–24 February 1997; pp. 21–34. [Google Scholar]
- Huang, Z. Extensions to the k-Means Algorithm for Clustering Large Data Sets with Categorical Values. Data Min. Knowl. Discov. 1998, 2, 283–304. [Google Scholar] [CrossRef]
- Cao, F.; Liang, J.; Bai, L. A new initialization method for categorical data clustering. Expert Syst. Appl. 2009, 36, 10223–10228. [Google Scholar] [CrossRef]
- Garbe, W. SymSpell. Available online: https://github.com/wolfgarbe/symspell (accessed on 5 February 2019).
- SymSpellpy. Available online: https://github.com/mammothb/symspellpy (accessed on 9 October 2019).
- Bird, S.; Klein, E.; Loper, E. Natural Language Processing with Python: Analyzing Text with the Natural Language Toolkit; O’Reilly Media, Inc.: Boston, MA, USA, 2009. [Google Scholar]
- Fellbaum, C. WordNet: An Electronic Lexical Database; MIT Press: Cambridge, MA, USA, 1998. [Google Scholar]
- Chen, T.; Li, M.; Li, Y.; Lin, M.; Wang, N.; Wang, M.; Xiao, T.; Xu, B.; Zhang, C.; Zhang, Z. MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems. In Proceedings of the 30th International Conference on Neural Information Processing Systems, Workshop on Machine Learning Systems (NIPS), Barcelona, Spain, 5–10 December 2016. [Google Scholar]
- Guo, J.; He, H.; He, T.; Lausen, L.; Li, M.; Lin, H.; Shi, X.; Wang, C.; Xie, J.; Zha, S.; et al. GluonCV and GluonNLP: Deep Learning in Computer Vision and Natural Language Processing. J. Mach. Learn. Res. 2020, 21, 1–7. [Google Scholar]
- McKinney, W. Data Structures for Statistical Computing in Python. In Proceedings of the 9th Python in Science Conference; van der Walt, S., Millman, J., Eds.; SciPy: Austin, TX, USA, 2010; pp. 56–61. [Google Scholar] [CrossRef] [Green Version]
- Virtanen, P.; Gommers, R.; Oliphant, T.E.; Haberland, M.; Reddy, T.; Cournapeau, D.; Burovski, E.; Peterson, P.; Weckesser, W.; Bright, J.; et al. SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python. Nat. Methods 2020, 17, 261–272. [Google Scholar] [CrossRef] [Green Version]
- Pedregosa, F.; Varoquaux, G.; Gramfort, A.; Michel, V.; Thirion, B.; Grisel, O.; Blondel, M.; Prettenhofer, P.; Weiss, R.; Dubourg, V.; et al. Scikit-learn: Machine Learning in Python. J. Mach. Learn. Res. 2011, 12, 2825–2830. [Google Scholar]
- Oliphant, T.E. A Guide to NumPy; Trelgol Publishing: Austin, TX, USA, 2006; Volume 1. [Google Scholar]
- Van der Walt, S.; Colbert, S.C.; Varoquaux, G. The NumPy Array: A Structure for Efficient Numerical Computation. Comput. Sci. Eng. 2011, 13, 22–30. [Google Scholar] [CrossRef] [Green Version]
- Krizhevsky, A. One Weird Trick for Parallelizing Convolutional Neural Networks; Technical Report; Google Inc.: Mountain View, CA, USA, 2014. [Google Scholar]
- Simonyan, K.; Zisserman, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. In Proceedings of the 3rd International Conference on Learning Representations (ICLR), San Diego, CA, USA, 7–9 May 2015. [Google Scholar]
- He, K.; Zhang, X.; Ren, S.; Sun, J. Identity Mappings in Deep Residual Networks. In Computer Vision—ECCV 2016; Springer: Berlin/Heidelberg, Germany, 2016; pp. 630–645. [Google Scholar]
- Chen, J.; Ngo, C.W. Deep-based ingredient recognition for cooking recipe retrieval. In Proceedings of the 24th ACM international conference on Multimedia; ACM: New York, NY, USA, 2016; pp. 32–41. [Google Scholar]
- Settles, B. From theories to queries: Active learning in practice. In Active Learning and Experimental Design Workshop in Conjunction with AISTATS 2010; Microtome Publishing: Brookline, MA, USA, 2011; pp. 1–18. [Google Scholar]
- Yanto, I.T.R.; Ismail, M.A.; Herawan, T. A modified Fuzzy k-Partition based on indiscernibility relation for categorical data clustering. Eng. Appl. Artif. Intell. 2016, 53, 41–52. [Google Scholar] [CrossRef]
- Li, C.; Jiang, L.; Xu, W. Noise correction to improve data and model quality for crowdsourcing. Eng. Appl. Artif. Intell. 2019, 82, 184–191. [Google Scholar] [CrossRef]
Category | WordNet Description |
---|---|
<site> | A computer connected to the internet that maintains a series of web pages on the World Wide Web. |
<library> | A building that houses a collection of books and other materials. |
<dunlin> | Small common sandpiper that breeds in northern or Arctic regions and winters in southern United States or Mediterranean regions |
<bolete> | Any fungus of the family Boletaceae |
<jacamar> | Tropical American insectivorous bird having a long sharp bill and iridescent green or bronze plumage |
<gyromitra> | Any fungus of the genus Gyromitra |
<dhole> | Fierce wild dog of the forests of central and southeast Asia that hunts in packs |
<kakatoe galerita> | White cockatoo with a yellow erectile crest |
<earthstar> | Any fungus of the family Geastraceae; in form suggesting a puffball whose outer peridium splits into the shape of a star. |
<siamang> | Large black gibbon of Sumatra having the 2nd and 3rd toes partially united by a web. |
No. | Category |
---|---|
1. | <pickles, shredded pork & vermicelli> |
2. | <fried lamb with cumin> |
3. | <four-joy meatballs> |
4. | <sauteed bullfrog with pickled peppers> |
5. | <saute spicy chicken> |
6. | <spare ribs with garlic> |
7. | <chicken feet with pickled peppers> |
8. | <sauteed shredded pork with skin of tofu> |
9. | <braised beef with brown sauce> |
10. | <sauteed shredded pork with garlic sprout> |
11. | <roast chicken wings> |
12. | <braised intestines in brown sauce> |
13. | <braised pork> |
14. | <sauteed snails> |
15. | <beefsteak> |
16. | <pork with garlic sauce> |
No. | Transformed Category | Original Categories |
---|---|---|
1. | <bird> | <dunlin>, <jacamar>, <kakatoe galerita> |
2. | <book> | <library> |
3. | <fox> | <dhole> |
4. | <library> | <library> |
5. | <unknown> | <library>, <bolete> |
6. | <monkey> | <siamang> |
7. | <mushroom> | <bolete>, <gyromitra>, <earthstar> |
8. | <website> | <site> |
Model | Top-1 ACC (%) | Precision (%) | Recall (%) | F1-Score (%) | ||||
---|---|---|---|---|---|---|---|---|
Pre | Post | Pre | Post | Pre | Post | Pre | Post | |
Human population | 99.38 | 99.38 | 90.91 | 96.88 | 90.34 | 99.74 | 90.62 | 98.08 |
AlexNet | 88.13 | 88.13 | 90.91 | 88.13 | 80.11 | 84.38 | 84.66 | 79.23 |
ResNet152v2 | 96.88 | 96.88 | 90.91 | 89.58 | 88.07 | 95.57 | 89.42 | 88.74 |
VGG19 | 94.38 | 94.38 | 90.91 | 88.75 | 85.8 | 91.92 | 88.08 | 85.26 |
Original Categories | Transformed Category | |||||||
---|---|---|---|---|---|---|---|---|
<meat> | <meatball> | <chicken foot> | <noodle> | <stew> | <snail> | <vegetable> | <fry stir> | |
<sauteed shredded pork with garlic sprout> | ✔ | ✔ | ✔ | |||||
<braised intestines in brown sauce> | ✔ | ✔ | ||||||
<fried lamb with cumin> | ✔ | ✔ | ||||||
<spare ribs with garlic> | ✔ | ✔ | ||||||
<roast chicken wings> | ✔ | |||||||
<beefsteak> | ✔ | |||||||
<braised pork> | ✔ | |||||||
<braised beef with brown sauce> | ✔ | ✔ | ✔ | |||||
<sauteed bullfrog with pickled peppers> | ✔ | ✔ | ✔ | ✔ | ||||
<pork with garlic sauce> | ✔ | ✔ | ✔ | ✔ | ||||
<four-joy meatballs> | ✔ | ✔ | ||||||
<sauteed shredded pork with skin of tofu> | ✔ | ✔ | ||||||
<pickles, shredded pork & vermicelli> | ✔ | ✔ | ✔ | |||||
<saute spicy chicken> | ✔ | ✔ | ✔ | |||||
<chicken feet with pickled peppers> | ✔ | |||||||
<sauteed snails> | ✔ | ✔ | ✔ | ✔ |
Original Categories | Transformed Category | ||||||
---|---|---|---|---|---|---|---|
<pork> | <meat> | <meatball> | <seafood> | <intestine> | <noodle> | <vegetable> | |
<braised intestines in brown sauce> | ✔ | ✔ | ✔ | ||||
<fried lamb with cumin> | ✔ | ✔ | ✔ | ✔ | |||
<spare ribs with garlic> | ✔ | ✔ | |||||
<roast chicken wings> | ✔ | ✔ | |||||
<beefsteak> | ✔ | ✔ | ✔ | ||||
<braised pork> | ✔ | ✔ | |||||
<braised beef with brown sauce> | ✔ | ✔ | ✔ | ||||
<sauteed bullfrog with pickled peppers> | ✔ | ✔ | ✔ | ||||
<pork with garlic sauce> | ✔ | ✔ | ✔ | ✔ | |||
<four-joy meatballs> | ✔ | ✔ | ✔ | ||||
<chicken feet with pickled peppers> | ✔ | ✔ | ✔ | ||||
<saute spicy chicken> | ✔ | ✔ | ✔ | ✔ | |||
<sauteed snails> | ✔ | ✔ | ✔ | ||||
<pickles, shredded pork & vermicelli> | ✔ | ✔ | |||||
<sauteed shredded pork with skin of tofu> | ✔ | ||||||
<sauteed shredded pork with garlic sprout> | ✔ |
Model | Top-1 ACC (%) | Precision (%) | Recall (%) | F1-Score (%) | ||||
---|---|---|---|---|---|---|---|---|
Pre | Post | Pre | Post | Pre | Post | Pre | Post | |
AlexNet | 89.93 | 90.63 | 93.11 | 88.79 | 84.64 | 81.4 | 88.42 | 84.86 |
Human | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 |
ResNet152v2 | 93.4 | 94.1 | 93.42 | 88.89 | 87.91 | 84.72 | 90.42 | 86.72 |
VGG19 | 93.4 | 94.44 | 93.07 | 88.89 | 87.91 | 83.9 | 90.33 | 86.27 |
Model | Top-1 ACC (%) | Precision (%) | Recall (%) | F1-Score (%) | ||||
---|---|---|---|---|---|---|---|---|
Pre | Post | Pre | Post | Pre | Post | Pre | Post | |
AlexNet | 89.93 | 90.63 | 93.11 | 87.21 | 84.64 | 80.23 | 88.42 | 83.54 |
Human | 100 | 100 | 100 | 100 | 100 | 100 | 100 | 100 |
ResNet152v2 | 93.4 | 94.1 | 93.42 | 87.5 | 87.91 | 83.47 | 90.42 | 85.41 |
VGG19 | 93.4 | 94.44 | 93.07 | 87.5 | 87.91 | 83.61 | 90.33 | 85.49 |
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Koporec, G.; Košir, A.; Leonardis, A.; Perš, J. Cognitive Relevance Transform for Population Re-Targeting. Sensors 2020, 20, 4668. https://doi.org/10.3390/s20174668
Koporec G, Košir A, Leonardis A, Perš J. Cognitive Relevance Transform for Population Re-Targeting. Sensors. 2020; 20(17):4668. https://doi.org/10.3390/s20174668
Chicago/Turabian StyleKoporec, Gregor, Andrej Košir, Aleš Leonardis, and Janez Perš. 2020. "Cognitive Relevance Transform for Population Re-Targeting" Sensors 20, no. 17: 4668. https://doi.org/10.3390/s20174668
APA StyleKoporec, G., Košir, A., Leonardis, A., & Perš, J. (2020). Cognitive Relevance Transform for Population Re-Targeting. Sensors, 20(17), 4668. https://doi.org/10.3390/s20174668