Relationship Between Human Brain and CNN

Hello, we all know that the image classification process of convolutional neural networks is influenced by the working principle of neural networks in the human brain. Let’s examine the relationship between them.
       Convolutional Neural Networks are deep learning architecture that is widely used in computer vision studies such as image classification, localization, object perception. Convolutional neural networks are the field of study associated with machine learning to analyzing visual imagery. CNN choose unique features from pictures to distinguish the given figure. This process also happening in our brains unconsciously.
Biological Inspiration of Convolutional Neural Network (CNN)
Mapping of human visual system and CNN
Research in Sensor Processing (1960’s and 1970’s)
These works are prime Dr. Hubel and Dr. Wiesel worked on the area of Sensory Processing. In which, they inserted a micro-electrode into the primary visual cortex of an partially anesthetized cat so that she can’t move and shown the images of line at different angles to the cat.
Through the micro-electrode they found that some neurons fired very rapidly by watching the lines at specific angles, while other neurons responded best to the lines at different angles. Some of these neurons responded to light and dark patterns differently, while other neurons responded to detect motion in the certain direction.

Where is visual cortex located in humans brain?


                                                                                     Figure 1: Functional areas for the human brain

Visual Cortex is the part of the Cerebral Cortex of the Brain that processes the visual information. Visual nerves from the eyes runs straight to the primary visual cortex. Based on the structural and the functional characteristics it is divided into different areas, as shown in the following picture:

Figure 2: Different areas of visual cortex

 Visual cortex: Functions
The visual information is passed from one cortical area to another and each cortical area is more specialized than the last one. The neurons in the specific field only respond to the specific actions.
Some of them with their functions are as follows:

  1. Primary visual cortex or V1: It preserves spatial location of visual information i.e. orientation of edges and lines. It is the first one to receive the signals form what eyes have captured.
  2. Secondary visual cortex or V2: It receives strong feed-forward connections from V1 and sends strong connections to V3, V4 and V5. It also sends strong feedback network to V1. Its function is to collects spatial frequency, size, color and shape of the object.
  3. Third visual cortex or V3: It receives inputs from V2. It helps in processing global motion and gives complete visual representation.

      4. V4: It also receives inputs from V2. It recognizes simple geometric shapes and also forms recognition of object. It is not tuned for complex objects              as Human Faces.

  1. Middle temporal (MT)visual area or V5: It is used to detect speed and direction of moving visual object i.e. motion perception. It also detects motion of complex visual features. It receives direct connections from V1.
  2. Dorsomedial (DM) area or V6: used to detect wide field and self-motion stimulation. Like V5 it also receives direct connections from V1. It has extremely sharp selection of the orientation of visual contours.

Structure of Convolutional Neural Networks
CNN processes the image with various layers.
Layers Of CNN
Input Layer: In this layer, data is transmitted raw to the network.
Convolutional Layer: Used for detecting features.
Non-Linearity Layer: Introduction to nonlinearity to the system
Pooling (Down sampling) Layer: Decrease count of weights and check the conformation
Flattening Layer: Prepare the data for classical neural network
Fully Connected Layer: Standard Neural Network used in classification

                  Figure 3: CovNet Diagram

As a result, CNN imitates the work of the visual cortex in our brain. If we look the plane picture, we can define the plane by separating the features such as two wings, engines, windows. CNN does the same thing but previously they detect low-level properties such as curves and edges and create them up to more abstract concepts. Don’t you think it’s great? Hope to see you in our next blog.

  4. Kuş, Zeki.”Mikrokanonikal Optimizasyon Algoritması ile Konvolüsyonel Sinir Ağlarında Hiper Parametrelerin Optimize Edilmesi”Fatih Sultan Mehmet University,2019 (pp. 16-21)


Leave a Reply