Effect of propionic acid on the rat’s behaviour: autism model

Lobzhanidze, Giorgi

doi:10.1016/j.euroneuro.2017.12.078

Cited by 2 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The latest neuroscience findings will be prioritized in rats for validation and implementation [1]- [3]. Many of the experiments used to study the effects of drugs on behavior were carried out in rats [4]- [6]. Some psychological experiments also use rats as samples [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Rat Behavior Observation System Based on Transfer Learning

Jin

Duan

2019

IEEE Access

View full text Add to dashboard Cite

Excellent rat behavior observation methods help promote scientific research in neuroscience, social sciences, and pharmacy. Almost all traditional rat behavior observation methods track rats in the fixed environment or through intrusive devices or markers, which may have an impact on rats. Recently, deep learning methods have achieved great success in the field of computer vision because of their powerful ability to feature extraction. However, it is disadvantageous that deep learning methods require a large number of labeled images as a training dataset to adjust its deep neural networks. In this paper, in order to apply the deep learning method to rat behavior observation, we adopted two transfer learning methods to reduce dataset and realized detecting rats in various environments without any intrusive devices or markers. In addition, with the help of track by detection method, we have completed the long-term tracking of multiple rats. We also proposed global category non-maximum suppression to classify rat postures accurately with deep neural networks, which provides researchers with more experimental attempts. In the observation of three rats for one hour, tracking identity definition only happens 34 times and the classification accuracy rate is 89.09%.INDEX TERMS Rats behavior observation, deep learning, transfer learning, tracking.

show abstract