ML Counting Task

A ML Counting Task is a Machine Learning Task for determining the number of elements of itemset or that of object instances.

• AKA: Counting Task, Learning to Count.
• Context:
  • It can be solved by a ML Counting System by implementing an ML Counting Algorithm.
• Example(s):
  • Leaf Counting Task,
  • Blood Cell Counting Task,
  • Bigram Counting Task,
  • Dynamic Itemset Counting (DIC) Task (Brin et al., 1997),
  • Triangle Counting Task,
• Counter-Example(s):
  • Adding Task,
  • Clustering Task.
• See: Combinatorics, Inclusive Counting, Finger Counting, Counting Function, Counting Measure.

---

---

References

2020

• (Wikipedia, 2020) ⇒ https://en.wikipedia.org/wiki/Counting Retrieved:2020-10-11.
  • Counting is the process of determining the number of elements of a finite set of objects. The traditional way of counting consists of continually increasing a (mental or spoken) counter by a unit for every element of the set, in some order, while marking (or displacing) those elements to avoid visiting the same element more than once, until no unmarked elements are left; if the counter was set to one after the first object, the value after visiting the final object gives the desired number of elements. The related term enumeration refers to uniquely identifying the elements of a finite (combinatorial) set or infinite set by assigning a number to each element.

    Counting sometimes involves numbers other than one; for example, when counting money, counting out change, "counting by twos" (2, 4, 6, 8, 10, 12, ...), or "counting by fives" (5, 10, 15, 20, 25, ...).

    There is archaeological evidence suggesting that humans have been counting for at least 50,000 years. Counting was primarily used by ancient cultures to keep track of social and economic data such as the number of group members, prey animals, property, or debts (that is, accountancy). Notched bones were also found in the Border Caves in South Africa that may suggest that the concept of counting was known to humans as far back as 44,000 BCE. The development of counting led to the development of mathematical notation, numeral systems, and writing.

2019

• (Alam & Islam, 2019) ⇒ Mohammad Mahmudul Alam, and Mohammad Tariqul Islam (2019). "Machine Learning Approach Of Automatic Identification And Counting Of Blood Cells". In: Healthcare technology letters, 6(4), 103-108.
  • QUOTE: In this Letter, a deep learning based blood cell counting method has been proposed. We employ a deep learning based object detection method to detect different blood cells. Among the state-of-the-arts object detection algorithms such as regions with convolutional neural network (R-CNN) (...)

2018

• (Ubbens et al., 2018) ⇒ Jordan Ubbens, Mikolaj Cieslak, Przemyslaw Prusinkiewicz, and Ian Stavness (2018). "The Use Of Plant Models In Deep Learning: An Application To Leaf Counting In Rosette Plants". In: Plant methods, 14(1), 6.
  • QUOTE: Deep learning refers to a broad category of machine learning techniques, which typically involve the learning of features in a hierarchical fashion. Such techniques have been shown to be successful in many types of computer vision tasks, including image classification, multi-instance detection, and segmentation. Deep learning is an area of active research, and applications to plant science are still in the early stages. Previous work has shown the advantage of deep learning in complex image-based plant phenotyping tasks over traditional hand-engineered computer vision pipelines for the same task. Such tasks include leaf counting, age estimation, mutant classification, plant disease detection and diagnosis from leaf images, the classification of fruits and other organs, as well as pixel-wise localization of root and shoot tips, and ears . The small body of existing research on deep learning applications in image-based plant phenotyping shows promise for future work in this field.

2015

• (Segui et al., 2015) ⇒ Santi Segui, Oriol Pujol, and Jordi Vitria (2015)."Learning To Count With Deep Object Features". In: Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPR Workshops 2015).
  • QUOTE: Learning to count is a learning strategy that has been recently proposed in the literature for dealing with problems where estimating the number of object instances in a scene is the final objective.
    (...)

     Counting the number of instances of an object in an image can be approached from two different perspectives: 1) training an object detector, and 2) training an object counter. In the first case we must provide the system with a large set of object examples, properly labeled and localized, that represent most of the possible views and appearances of the object, and the result is an object classifier. In the second case we only need to provide the number of object instances for each image sample and the result is typically a regressor (...)

1997

• (Brin et al., 1997) ⇒ Sergey Brin, Rajeev Motwani, Jeffrey D. Ullman, and Shalom Tsur (1997, June). "Dynamic Itemset Counting And Implication Rules For Market Basket Data". In: Proceedings of the 1997 ACM SIGMOD International Conference on Management of data (pp. 255-264).
  • QUOTE: An algorithm which counts all the large itemsets must find and count all of the large itemsets and the minimal small itemsets (that is, all of the boxes and circles). The DIC algorithm, described here, marks itemsets in four different possible ways:
    • Solid box - confirmed large itemset - an itemset we have finished counting that exceeds the support threshold.
    • Solid circle - confirmed small itemset - an itemset we have finished counting that is below the support threshold.
    • Dashed box - suspected large itemset - an itemset we are still counting that exceeds the support threshold.
    • Dashed circle - suspected small itemset - an itemset we are still counting that is below the support threshold.

      (...)

---