Fly Feature Detectors Show Contrast Invariance, Omni-Directionality, Velocity Constancy, and Octopaminergic Loss of Background Motion Suppression

This chapter provides an overview of biological perceptual systems and their underlying computational principles focusing on the sensory sheets of the retina and cochlea and exploring how complex feature detection emerges by combining simple feature detectors in a hierarchical fashion. We also explore how the microcircuits of the neocortex implement such schemes pointing out similarities to progress in the field of machine vision driven deep learning algorithms. We see signs that engineered systems are catching up with the brain. For example, vision-based pedestrian detection systems are now accurate enough to be installed as safety devices in (for now) human-driven vehicles and the speech recognition systems embedded in smartphones have become increasingly impressive. While not being entirely biologically based, we note that computational neuroscience, as described in this chapter, makes up a considerable portion of such systems’ intellectual pedigree.

Download Full-text

Frequency Variability Feature for Life Signs Detection and Localization in Natural Disasters

Remote Sensing ◽

10.3390/rs13040796 ◽

2021 ◽

Vol 13 (4) ◽

pp. 796

Author(s):

Long Zhang ◽

Xuezhi Yang ◽

Jing Shen

Keyword(s):

Natural Disasters ◽

Current Method ◽

Peak Frequency ◽

Significant Information ◽

Physiological Phenomenon ◽

In The Wild ◽

Motion Magnification ◽

Background Motion ◽

Detection And Localization ◽

Frequency Properties

The locations and breathing signal of people in disaster areas are significant information for search and rescue missions in prioritizing operations to save more lives. For detecting the living people who are lying on the ground and covered with dust, debris or ashes, a motion magnification-based method has recently been proposed. This current method estimates the locations and breathing signal of people from a drone video by assuming that only human breathing-related motions exist in the video. However, in natural disasters, background motions, such as swing trees and grass caused by wind, are mixed with human breathing, that distort this assumption, resulting in misleading or even no life signs locations. Therefore, the life signs in disaster areas are challenging to be detected due to the undesired background motions. Note that human breathing is a natural physiological phenomenon, and it is a periodic motion with a steady peak frequency; while background motion always involves complex space-time behaviors, their peak frequencies seem to be variable over time. Therefore, in this work we analyze and focus on the frequency properties of motions to model a frequency variability feature used for extracting only human breathing, while eliminating irrelevant background motions in the video, which would ease the challenge in detection and localization of life signs. The proposed method was validated with both drone and camera videos recorded in the wild. The average precision measures of our method for drone and camera videos were 0.94 and 0.92, which are higher than that of compared methods, demonstrating that our method is more robust and accurate to background motions. The implications and limitations regarding the frequency variability feature were discussed.

Download Full-text

Unsupervised learning by competing hidden units

Proceedings of the National Academy of Sciences ◽

10.1073/pnas.1820458116 ◽

2019 ◽

Vol 116 (16) ◽

pp. 7723-7731 ◽

Cited By ~ 16

Author(s):

Dmitry Krotov ◽

John J. Hopfield

Keyword(s):

Learning Algorithm ◽

Lower Layer ◽

Learning Rule ◽

Backpropagation Algorithm ◽

Feedforward Networks ◽

Feature Detectors ◽

End To End ◽

Hidden Layer ◽

Full Network ◽

Global Inhibition

It is widely believed that end-to-end training with the backpropagation algorithm is essential for learning good feature detectors in early layers of artificial neural networks, so that these detectors are useful for the task performed by the higher layers of that neural network. At the same time, the traditional form of backpropagation is biologically implausible. In the present paper we propose an unusual learning rule, which has a degree of biological plausibility and which is motivated by Hebb’s idea that change of the synapse strength should be local—i.e., should depend only on the activities of the pre- and postsynaptic neurons. We design a learning algorithm that utilizes global inhibition in the hidden layer and is capable of learning early feature detectors in a completely unsupervised way. These learned lower-layer feature detectors can be used to train higher-layer weights in a usual supervised way so that the performance of the full network is comparable to the performance of standard feedforward networks trained end-to-end with a backpropagation algorithm on simple tasks.

Download Full-text