Sunday, January 17, 2010

Table of results for MNIST dataset

This is a table documenting some of the best results some paper obtained in MNIST dataset.

Results shown indicates the error obtained by training on all 60,000 samples and testing on 10,000 samples.
  1. Multi-column Deep Neural Networks for Image Classification (CVPR 2012)
    Cited 9 times. 0.23%

    Supplemental material, Technical Report
  2. Deep Big Simple Neural Nets Excel on Handwritten Digit Recognition (2010)
    Cited 1 time. 0.35%
    Additional info: 6-layer NN 784-2500-2000-1500-1000-500-10 (on GPU) [elastic distortions]
  3. Efficient Learning of Sparse Representations with an Energy-Based Model (2006)
    Cited 109 times. 0.39%
    Additional info: large conv. net, unsup pretraining [elastic distortions]
  4. Stochastic Pooling for Regularization of Deep Convolutional Neural Networks (2013)
    Cited 1 times. 0.47%
    Additional info: Stochastic Pooling
  5. Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis (2003)
    Cited 190 times. 0.4%
  6. What is the Best Multi-Stage Architecture for Object Recognition? (ICCV 2009)
    Cited 39 times. 0.53%
    Additional info: large conv. net, unsup pretraining [no distortions]
  7. Deformation Models for Image Recognition (PAMI 2007)
    Cited 46 times. 0.54%
    Additional info: K-NN with non-linear deformation (IDM) (Preprocessing: shiftable edges)
  8. A trainable feature extractor for handwritten digit recognition (2007)
    Cited 38 times. 0.54%
    Additional info: Trainable feature extractor + SVMs [affine distortions]
  9. Training Invariant Support Vector Machines (2002)
    Cited 281 times. 0.56%
    Additional info: Virtual SVM, deg-9 poly, 2-pixel jittered (Preprocessing: deskewing)
  10. Simple Methods for High-Performance Digit Recognition Based on Sparse Coding (TNN 2008)
    0.59%
    Additional info: unsupervised sparse features + SVM, [no distortions]
  11. Unsupervised learning of invariant feature hierarchies with applications to object recognition (CVPR 2007)
    Cited 119 times. 0.62%
    Additional info: large conv. net, unsup features [no distortions]
  12. Shape matching and object recognition using shape contexts (PAMI 2002)
    Cited 2089 times. 0.63%
    Additional info: K-NN, shape context matching (preprocessing: shape context feature extraction)
  13. Beyond Spatial Pyramids: Receptive Field Learning for Pooled Image Features (2012)
    Cited 0 times. 0.64%
  14. Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representations (2009)
    0.82%
  15. Large-Margin kNN Classification using a Deep Encoder Network (2009)
    0.94%
  16. Deep Boltzmann Machines (2009)
    0.95%
  17. CS81: Learning words with Deep Belief Networks (2008)
    1.12%
  18. Convolutional Neural Networks (2003)
    1.19%
    More info: The ConvNN is based on the paper "Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis".
  19. Reducing the dimensionality of data with neural networks (2006)
    1.2%
  20. Deep learning via semi-supervised embedding (2008)
    1.5%

Sunday, October 25, 2009

Table of results for Caltech 256 dataset

This is a table documenting some of the best results some paper obtained in Caltech-256 dataset.


Results shown here are all trained using 30 samples from each category.
  1. Visualizing and Understanding Convolutional Networks (ARXIV 2013)
    Cited 14 times. 70.6% ± 0.2%
  2. Multipath Sparse Coding Using Hierarchical Matching Pursuit (CVPR 2013)
    Cited 7 times. 50.7%
    Additional info: Multipath Hierarchical Matching Pursuit
    Link to paper's project page
  3. Learning Subcategory Relevances for Category Recognition (CVPR 2008)
    Cited 48 times. 49.5%
  4. Spatially Local Coding for Object Recognition (ACCV 2010)
    Cited 1 time. 46.6% ± 0.2%

    Additional info: Multi-scale SIFT features extracted every 4 pixels.

    Link to paper's project page
    Link to paper's source code
  5. On Feature Combination for Multiclass Object Detection (ICCV 2009)
    Cited 376 times. 45.8%
    Additional info:
    LP-β
    Link to paper's project page (Contains results, source code and pre-computed features)
  6. Image Classification using Random Forests and Ferns (2007)
    Cited 412 times. 45.3%
  7. Local Pyramidal Descriptors for Image Recognition (PAMI 2013)
    Cited 1 time. 44.86%
    Additional info: 
    P-SIFT + Fisher encoding + SPM + Linear SVM
    Link to paper's project page (Contains source code and demo)
  8. A Binary Classification Framework for Two-Stage Multiple Kernel Learning (2012)
    Cited 5 times. 44.8%
  9. Efficient Learning of Sparse, Distributed, Convolutional Feature Representations for Object Recognition (ICCV 2011)
    Cited 21 times. 42.05%
    Additional info: CRBM K=4096
  10. In Defense of Nearest-Neighbor Based Image Classification (CVPR 2008)
    Cited 478 times. 42%
    Additional info: NBNN (5 descriptors)
  11. Locality-constrained Linear Coding for Image Classification (CVPR 2010)
    Cited 547 times. 41.19%
  12. Local Naive Bayes Nearest Neighbor for Image Classification (2011)
    Cited 20 times. 40.1%
  13. Sparse Spatial Coding: A Novel Approach for Efficient and Accurate Object Recognition (ICRA 2012)
    Cited 9 times. 37.08% ± 0.36%
  14. Caltech-256 object categoriy dataset (2007)
    Cited 596 times. 34.1%
  15. Linear spatial pyramid matching using sparse coding for image classification (CVPR 2009)
    Cited 713 times. 34.02%
  16. Kernel codebooks for scene categorization (ECCV 2008)
    Cited 242 times. 27.17%

Friday, August 21, 2009

Table of results for Caltech 101 dataset

This is a table documenting some of the best results some paper obtained in Caltech-101 dataset.

Results shown here are all trained using 30 samples from each category.
  1. Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition (ARXIV 2014)
    Cited 191 times. 91.44%
     ± 0.7%
    Link to paper's project page
  2. Visualizing and Understanding Convolutional Networks (ARXIV 2013)
    Cited 561 times. 86.5% ± 0.5%
    Additional info: Pre-trained with ImageNet
  3. Group-Sensitive Multiple Kernel Learning for Object Categorization (ICCV 2009)
    Cited 127 times. 84.3%
    Additional Info: GS-MKL
  4. Reference-Based Scheme Combined With K-SVD for Scene Image Categorization (2013)
    Cited 2 times. 83%
  5. Multipath Sparse Coding Using Hierarchical Matching Pursuit (CVPR 2013)
    Cited 2 times. 82.5%
     ± 0.5%
    Additional Info: Multipath Hierarchical Matching Pursuit
    Link to paper's project page
  6. LP-Beta + Geometric blur + PHOW gray/color + Self-Similarity
    82.1% ± 0.3%
  7. Learning Subcategory Relevances for Category Recognition (CVPR 2008)
    Cited 40 times. 81.9%
  8. Object Recognition as Ranking Holistic Figure-Ground Hypotheses (CVPR 2010)
    Cited 68 times. 81.9%
    Additional Info: Regression with Post-Processing.
  9. Image Classification using Random Forests and Ferns (ICCV 2007)
    Cited 378 times. 81.3%
    Additional Info: Bosch Multi-way SVM
  10. Spatially Local Coding for Object Recognition (ACCV 2010)
    Cited 0 time. 81% 
    ± 0.2%
    Additional Info: Multi-scale SIFT features extracted every 4 pixels.
    Link to paper's project page
    Link to paper's source code
    Link to paper's poster
  11. Local Pyramidal Descriptors for Image Recognition (PAMI 2013)
    Cited 1 time. 80.13%
    Additional Info: P-SIFT + Fisher encoding + SPM + Linear SVM
    Link to paper's project page (Contains source code, demo available)
  12. Sparse Spatial Coding: A Novel Approach for Efficient and Accurate Object Recognition (ICRA 2012)
    Cited 7 times. 80.02% ± 0.36%
    Additional Info: dictionary size is 4096.
  13. Distance-based Mixture Modeling for Classification via Hypothetical Local Mapping (2013)
    Cited 0 times. 80%
     ± 0.75%
  14. In Defense of Nearest-Neighbor Based Image Classification (CVPR 2008)
    Cited 442 times. 79.23%
    Additional Info: NBNN (5 descriptors)
  15. Smooth Sparse Coding via Marginal Regression for Learning Sparse Representations (2012)
    Cited 0 times. 79.11
     ± 0.87%
    Additional info: dictionary size = 4096
  16. Unsupervised and Supervised Visual Codes with Restricted Boltzmann Machines (ECCV 2012)
    Cited 4 times. 78.9% ± 1.1%
    Additional info: 1024 codewords trained on macrofeatures with supervised fine-tuning.
    Link to paper's poster
  17. Robust Classification of Objects, Faces, and Flowers Using Natural Image Statistics (CVPR 2010)
    Cited 49 times. 78.5% ± 0.5%
    Link to paper's supplemental material
    Link to paper's project page
    Link to paper's source code (MATLAB)
  18. Visual Geometric Group (VGG)'s implementation of Multiple Kernel Image Classifier trained on dense SIFT, self-similarity, and geometric blur features
    78.20% ± 0.4%
    Additional Info: Result of 77.8% is obtained by combining dense SIFT, self-similarity, and geometric blur features with the multiple kernel learning
  19. Efficient Learning of Sparse, Distributed, Convolutional Feature Representations for Object Recognition (ICCV 2011)
    Cited 19 times. 77.8%
    Additional info: CRBM K=4096
  20. On Feature Combination for Multiclass Object Detection (ICCV 2009)
    Cited 312 times. 77.8% ± 0.4%
    Additional info: LP-β
    Link to paper's project page (Contains results, source code and pre-computed features)
  21. Representing shape with a spatial pyramid kernel (CIVR 2007)Cited 547 times. 77.8%
  22. Additional info: Result of 77.8% is obtained by combining all 4 cues (shape 180, shape 360, gray appearance and color appearance.
  23. The devil is in the details - an evaluation of recent feature encoding methods (BMVC 2011)
    Cited 91 times. 77.78% ± 0.56%
    Additional info: Fisher (FK)
    Link to paper's project page (Contains dataset and source code)
  24. Object Recognition with Hierarchical Kernel Descriptors (CVPR 2011)
    Cited 29 times. 77.5%
    Link to paper's project page (Contains dataset, demos and source code)
  25. Ask the locals: multi-way local pooling for image recognition (ICCV 2011)
    Cited 44 times. 77.3%
     ± 0.6%
    Link to paper's supplemental material
  26. A Binary Classification Framework for Two-Stage Multiple Kernel Learning (ICML 2012)
    Cited 5 times. 77.2%
  27. Kernel Descriptors for Visual Recognition (NIPS 2010)
    Cited 48 times. 76.4% ± 0.7%
    Additional info: KDES-A(M)
  28. Local Naive Bayes Nearest Neighbor for Image Classification (CVPR 2012)
    Cited 17 times. 76% 
    ± 0.9%
    Link to paper's associated technical report
    Link to paper's source code
    Link to paper's project page
  29. Fast approximations to structured sparse coding and applications to object classification (2012)
    Cited 2 times. 75.7% ± 1%
  30. Learning mid-level features for recognition (CVPR 2010)
    Cited 221 times. 75.7% ± 1.1%
    Additional Info: Sparse Codes, Intersection Kernel.
  31. Object and Action Classification with Latent Window Parameters (IJCV 2013)
    Cited 0 times. 75.31% 
    ± 0.68%
  32. Beyond Spatial Pyramids: Receptive Field Learning for Pooled Image Features (2012)
    Cited 27 times. 75.3% ± 0.7%
  33. Image classification with multiple feature (2011)
    Cited 3 times. 75% ± 0.8%
  34. In Defense of Soft-assignment Coding (ICCV 2011)
    Cited 46 times. 74.2% ± 0.8%
    Link to author's web site
    Link to paper's source code
  35. Locality-constrained Linear Coding for Image Classification (CVPR 2010)
    Cited 446 times. 73.44%
    Project web site: Link to Project web site
    Source code: Link to MATLAB code (rar)
  36. Linear Spatial Pyramid Matching Using Sparse Coding for Image Classification (CVPR 2009)
    Cited 621 times. 73.2% ± 0.54%
    Additional Info: Sparse coding, max pooling, linear SVM
    Project web site: Link to Project web site
    Source code: Link to MATLAB code (rar)
  37. High Dimensional Nonlinear Learning using Local Coordinate Coding (Technical Report 2009)
    Cited 3 times. 73.14%
    Additional Info: Local coordinate coding, max pooling, linear SVM
  38. Recognition using Regions (CVPR 2009)
    Cited 156 times. 73.1%
  39. The importance of Encoding Versus Training with Sparse Coding and Vector Quantization (ICML 2011)
    Cited 101 times. 72.6%
  40. Learning Coupled Conditional Random Field for Image Decomposition with Application on Object Categorization (CVPR 2008)
    Cited 10 times. 70.38%
  41. Fast Image Search for Learned Metrics (CVPR 2008)
    Cited 126 times. 69.6%
    Additional info: ML+CORR
  42. A Multi-Scale Learning Framework for Visual Categorization (ACCV 2010)
    Cited 2 times. 68.5%
    Additional info: sparse coding (K = 900)
  43. Caltech-256 Object Category Dataset (2007)
    Cited 544 times. 67.6%
    Additional Info: Griffin's SPM
  44. Improved Spatial Pyramid Matching for Image Classification (ACCV 2010)
    Cited 3 times. 67.36% ± 0.17%
  45. Variable Sparsity Kernel Learning (JMLR 2011)
    Cited 23 times. 67.07%
  46. Deep Learning of Invariant Features via Simulated Fixations in Video (2012)
    Cited 2 times. 
    66%
    Additional info: Trained also with 
    video (unrelated to Caltech-101) obtained 74.6%
  47. Similarity-based cross-layered hierarchical representations for object categorization (CVPR 2008)
    Cited 47 times. 66.5%
    Additional Info: Shapinals
  48. SVM-KNN - Discriminative Nearest Neighbor Classification for Visual Category Recognition (CVPR 2006)
    Cited 568 times. 66.2% ± 0.5%
  49. The Hierarchical Beta Process for Convolutional Factor Analysis and Deep Learning (ICML 2011)
    Cited 7 times. 65.8% ± 0.6%
  50. Combined Descriptors in Spatial Pyramid Domain for Image Classification (2012)
    Cited 0 times. 65.5% ± 0.49%
  51. Bag-of-Features Kernel Eigen Spaces for Classification (ICPR 2008)
    Cited 2 times. 65.5% ± 0.7%
  52. Image Retrieval and Classification using Local Distance Functions (NIPS 2006)
    Cited 136 times. 65.2%
  53. Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representations (ICML 2009)
    Cited 315 times. 65.4% ± 0.5%
    Additional Info: CDBN (first+second layers)
  54. Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories (CVPR 2006)
    Cited 2582 times. 64.6% ± 0.8%
    Additional Info: L=2, M=200, Pyramid
    S
    ource code: Link to MATLAB code (zip)
    Slides: Link
  55. Kernel Codebooks for Scene Categorization (ECCV 2008)
    Cited 224 times. 64.12%
  56. Visual Word Ambiguity (PAMI 2010)
    Cited 257 times. 64.1%
  57. Using dependent regions for object categorization in a generative framework (CVPR 2006)
    Cited 121 times. 63%
  58. SIFTing the Relevant from the Irrelevant - Automatically Detecting Objects in Training Images (2009)
    Cited 0 times. 61.45%
  59. Pyramid Match Kernels: Discriminative Classification with Sets of Image Features (ICCV 2005)
    Cited 849 times. 58.2%
  60. Efficient Classification for Additive Kernel SVMs (PAMI 2012)
    Cited 11 times. 56.59% ± 0.77%
  61. Max-Margin Additive Classifiers for Detection (ICCV 2009)
    Cited 79 times. 56.49%
  62. Multiclass Object Recognition with Sparse, Localized Features (CVPR 2006)
    Cited 345 times. 56%
  63. Efficiently Matching Sets of Features with Random Histograms (2008)
    Cited 46 times. 54.1%
  64. Unsupervised Learning of Invariant Feature Hierarchies with Applications to Object Recognition (2007)
    Cited 235 times. 54%
  65. Fast Inference in Sparse Coding Algorithms with Applications to Object Recognition (2008)
    Cited 64 times. 53%
  66. Classification using Intersection Kernel Support Vector Machines is Efficient (2008)
    Cited 396 times. 52%
    Project web site: Link
    Source code: Link to MATLAB/C code (tar.gz)
  67. Object Recognition with Features Inspired by Visual Cortex (2006)
    Cited 536 times. 42%

Sunday, April 05, 2009

Re-focused

In an attempt to re-focuses my energy on IT-related stuff that I will be doing, I have been doing some thinking and came up with a list (The list is not final):

Research
(I did not say machine learning, nor pattern recognition, as those are just means to an end)
- Computer Vision
- Natural Language Processing

Development
(I did not say functional programming, as those are just means to an end)
- Writing parallel & concurrent programs (Parallel processing, Concurrent programming)
- Writing maintainable, beautiful code

Design
- Designing simple, practical & nice UI

Friday, March 06, 2009

Return to research

This is a post foreshadowing my come back to the research world. I've been thinking a lot about vision. How do we humans perceive things? How do we recognize things? How do we get a sense of deja vu after seeing things we have seen before?

Even though I do not have the answer to these perplexing questions. I do however think that I *might* have a solution that might solve it to a certain degree. That's a pretty modest statement. But I'm not going to claim anything more exotic before I come up with a working prototype.

The ideas I have in mind have a lot to do with time-based learning, that is, learning not just things but the relationships they have. That is, that certain meaningful things appear to have an order. For example, you do not see a cat walking by, then suddenly it become a dog for half a second, then magically becoming a cat again. Things are not random. They appear in logical order. The idea is nothing new. It has been addressed by Jeff Hawkins in his work on Hierarchical Temporal Memory.

End of part 1. I will talk more about this later.

Monday, December 29, 2008

Programming skill...

Programming skill...
  • Those that has it, talks about why there are people who just can't program (no matter what).
  • Those that don't have it, talks about why the world isn't fair.
Why do I say this? Because if I see another article/blog that says some people just can't program no matter what, I need to pull his head off.

Disclaimer: In no way did I claim I'm superior in programming skills.

Sunday, June 08, 2008

Why good is actually bad.

Basic notion has it that as we gain more experience, our peers or children gets to reap the benefits. Well, the problem is, that is where the problem lies. You see, human learns by experience, be it good or bad. When we made a mistake, some external signals, be it your teacher or your mum tells you not to do it again (Bad experience). When we did something right, you are rewarded or praised (Good experience). Call it reinforcement learning, call it supervised learning, the point is, if you did something wrong, if the signs that you did something wrong was hidden from you, you are learning to repeat your wrongdoing. You may not implied that, or want to, but believe it or not, you are (asking him to repeat the wrong doing). There is no such thing as, "You are doing it wrong, but I'll just let you go this time.". It makes absolutely no sense. If you look at this way, this is common sense. We may not like being scold or make mistakes, it's all aweful medicine, but sometimes the patient just needed it.

The scary thing is, that’s what we do every now and then. Let me tell of a fictional story to illustrate the problem. A software programmer discovered that with C++, it is difficult to write code that does not leak memory, he suggested the team to use Python instead, before the team has decide which language to use. All is good and well. Well, the problem is C++ is still prevalent. Many legacy systems are still written in C/C++. When one of the other programmers in the team was assigned to maintain legacy software written in C++, he might not perform. Why? He/She has never learned best practices of C++! And the sole reason of that is because he/she never gets to learn it anyway; every time the chances come one of the programmer in the team will recommend the team to use a more modern, safer language. So you see, the point is, if you are only told that it's a wrong way and you should not go there, you think that it has saves you whole lot of trouble, but the truly sad truth is, it does not mean that you have actually learn anything. You merely avoided the problem.

Now, some of you might argue that there is simply too much to learn, it is unfeasible to learn everything. While that is true, if that something to learn is something fundamental, it is a mistake not to go through the entire learning process. This is the case where shortcut is not a good idea. Like the famous saying goes, “If we do not learn from history, we are bound to repeat it”. And the only way to learn history is to be in history. Walk in their shoes. Doing what they are doing. Merely being told the mistakes of the past does nothing to prevent it from happening it again, since there may be new, unexplored alternative path that will ultimately lead to the same mistake. And the only way to prevent it from happening is to have greater understanding of the nature of the mistake, and learning how to prevent it from ever happening again.

Perhaps, this is why there has been talk that the quality of Computer Science graduates are dropping, due to the fact that Java was chosen as most CS first language, where as lower level languages like assembly and C/C++ should be taught first, because they not only taught the students about the language itself, but also the fundamentals of computer architecture itself, how things work in the low level and etc. You can read all about it in [Computer Science Education: Where Are the Software Engineers of Tomorrow?].

The truth is, it is extremely dangerous to say that all history is useless experience, though that is not to say that there is no such thing as useless experiences. If you studied machine learning, you will know that, a typical artificial neural network would become brain dead if it is only showed positive samples without any negative samples during the training phase. A human brain works almost in the same principle, since a human brain is really both a discriminative and a generative system. Throwing off the discriminative part of our brain and it is likely that we will never learn. Similarly, if a baby is only taught the right way of doing things, the baby will not have a clue of if what he is doing is right or wrong. In other words, his discriminative ability, his ability to discern what is right or wrong, what is good or bad is simply not there or severely random and thus flawed.

Even if there are times when we needed to choose which knowledge to gain, I shudder to think at the outcome of making the wrong judgement on which knowledge is useless, and which is not: Because depending on what the situation is, it's actually very hard for us to really know for sure. Like Steve Jobs said, it's impossible for us to connect to dots ahead of time. We can only do so, looking back.

This is all obvious and understandable. However, recently, it is increasingly becoming harder to made mistakes, as we are becoming more intolerant to mistakes, while technology is becoming more adept at hiding complexities that are deemed to cause human mistakes. In the office, you are not allowed to use tools that your project leader deemed is bad, you will be fired for making the wrong decisions; In the university, you can't opt for subjects that are deemed to be obsolete by some, you are forced to answer questions based on what is stated on the textbook instead of your own thinking, which ultimately affected our creativity as mentioned in this article; In the house, children are taught not to talk to strangers, just like what Bruce Scheneier in his article "The Kindness of Strangers". In almost every cases, one is simply not acceptable to be different from others, you have to be just like everyone else, even though everyone else could be wrong. We should be given guidance, yes, but more importantly, we should be given the chance to make mistakes! Not repeating the same mistakes should be everyone's responsibility. Not the enforcer. Ideally.

To summarize, discriminative capability is important for us to learn. It is extremely crucial that we are left to made mistakes so that we would not do it again. The experience of doing something wrong is much more important than the experience of doing something correctly. Now, if there is an alternative way for us to experience the whole crucial process of making mistakes without actually wasting any time for it and yet learn a good lesson out of it, I would really like to know about it.

Monday, May 05, 2008

End of examination. Begining of research life.

Today marks the last day of my examination in this 3 year UTAR life. And I feel like this is the worst examination of all, and the only paper that I have the confidence to get at least A- is AI. To me, doing good in examination just don't motivate me that much anymore, what motivates me more now is research and development in things related to applied AI and computer vision. Which although sounds hard, is actually a lot easier than you might think, because there is no fixed and established way of doing things. Therefore, it tends to tax your creativity more. It's a lot harder to study, understand and memorizes things that other genius scientist/engineer has established, which is what is expected of you in university. In world of research, you don't necessary have to do it using other scientists' ways. They could be wrong and there seems to be always room for improvement. Everyone is solving a small piece of the puzzle, and no one knows the complete solution.

However, before I officially begin my research life, I will take a detour to work in Panasonic R&D for 1 year or so, commencing on June, before coming back to UTAR to begin my Master degree course.

Friday, February 01, 2008

My AI mini project

UCEC3064 - Intelligent Technique for Engineering Applications

Proposed mini project title:
Improving Circular Pairwise Neural Network Performance on Multiclass Classification Problems

Abstract:
Traditionally, neural networks are trained in a monolithic fashion. That is, a neural network would be trained to classify all K classes, preferably given equivalent amount of training samples for each classes. Training is slow and is unable to take advantage of current multi-cores or multi-processor (SMP) systems to train different classes concurrently. The research done by Teo Choon Hui on circular pairwise classification is an attempt to remedy this by breaking down a K class problem into k binary circular pairwise classification sub-problems. However, this method will reduce recognition accuracy due to the fact that there is a lacked of direct competition between certain pairs of classes. But, at the same time, such a method reduces training time by almost a factor of 3. In this research, we will compare the results of a binary one-versus-all classifier and a pairwise classifier with a circular pairwise classifier as proposed by Teo Choon Hui and attempts to improve recognition accuracy by experimenting with selective circular pairwise classification, in which hard binary sub-problems are paired instead of randomly choosing any 2 classes. We will then attempts to introduce an N-th class into each binary classifier to become a ternary circular pairwise classifier and evaluate its performance.

Keywords:
One-versus-all, Binary classification, Pairwise, Single-winner election methods, Circular Pairwise