Using classifiers as heuristics to describe local structure in Active Shape Models with small training sets

Tedin, Rafael; Becerra, J. A.; Duro, Richard J.

doi:10.1016/j.patrec.2013.04.026

Cited by 3 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is also important to note that the search method does not really need high precision levels in the estimation of the true position of the landmark. In Tedín et al [9] the authors proposed the Heuristic Local Appearance Model (HLAM). They successfully applied it for a small training set of hands under varying illumination and for a synthetic training set where the images presented spurious borders that fooled other local appearance models.…”

Section: A Active Shape Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Computational Intelligence based construction of a Body Condition Assessment system for cattle

Tedin

Becerra

Duro

et al. 2013

2013 IEEE International Conference on Computational Intelligence and Virtual Environments for Measurement Systems and Applicati

Self Cite

View full text Add to dashboard Cite

The objective of this paper is to describe a Computational Intelligence based Automatic Body Conditioning System for cattle we have called Automatic Body Condition Assessment (ABiCA). It is an automatic body condition scoring system for dairy cattle that aims to overcome the flaws of the subjective and time consuming scoring task that is usually carried out by experts. No special set-ups are needed since the system uses pictures taken using normal hand-held cameras. ABiCA is split into two components. A first component for the segmentation of the rear-end shape of a cow from its picture through Active Shape Models Active Shape Models (ASMs) that are evolved using an evolutionary algorithm. The second component is in charge of estimating the Body Condition Score (BCS) of a cow from the shape provided by the ASM. Several classifiers and a symbolic regression function evolved by means of genetic programming techniques are tested for this task. The whole system is tested over a set of images coming from different cattle farms and its goodness provided in terms of the classifications obtained by a set of experts.

show abstract

Section: A Active Shape Modelsmentioning

confidence: 99%

“…Here the same classifier as in Tedín et al [9] was chosen: the M5' classifier [10]. This classifier is implemented in the well-known WEKA [11] data mining software distribution, and this is the implementation used here.…”

Section: A Active Shape Modelsmentioning

confidence: 99%