Abstract:The Hubel–Wiesel (HW) model is a classical neurobiological model for explaining the orientation selectivity of cortical cells. However, the HW model still has not been fully proved physiologically, and there are few concise but efficient systems to quantify and simulate the HW model and can be used for object orientation detection applications. To realize a straightforward and efficient quantitive method and validate the HW model’s reasonability and practicality, we use McCulloch-Pitts (MP) neuron model to simulate simple cells and complex cells and implement an artificial visual system (AVS) for two-dimensional object orientation detection. First, we realize four types of simple cells that are only responsible for detecting a specific orientation angle locally. Complex cells are realized with the sum function. Every local orientation information of an object is collected by simple cells and subsequently converged to the corresponding same type complex cells for computing global activation degree. Finally, the global orientation is obtained according to the activation degree of each type of complex cell. Based on this scheme, an AVS for global orientation detection is constructed. We conducted computer simulations to prove the feasibility and effectiveness of our scheme and the AVS. Computer simulations show that the mechanism-based AVS can make accurate orientation discrimination and shows striking biological similarities with the natural visual system, which indirectly proves the rationality of the Hubel–Wiesel model. Furthermore, compared with traditional CNN, we find that our AVS beats CNN on orientation detection tasks in identification accuracy, noise resistance, computation and learning cost, hardware implementation, and reasonability.

Automatic Acquisition Algorithm of Orientation Words Based on Neural Network

Research on Text Orientation Analysis System Based on Neural Network

Arbitrarily-Oriented Text Recognition.

Exploring Neural Architecture Search for Text Classification

Deep learning based language and orientation recognition in document analysis

AON: Towards Arbitrarily-Oriented Text Recognition

Detecting texts of arbitrary orientations in natural images

MASTER: Multi-Aspect Non-local Network for Scene Text Recognition

Rotation-invariant features for multi-oriented text detection in natural images

A Novel Robust Algorithm for Position and Orientation Detection Based on Cascaded Deep Neural Network

MA-CharNet: Multi-angle fusion character recognition network

Multi-oriented Text Detection from Natural Scene Images Based on a CNN and Pruning Non-Adjacent Graph Edges

Automatic Photo Orientation Detection with Convolutional Neural Networks

Orientation Analysis of Web Review Based on Template Matching

Research on algorithm for networks new words identification

GRA: Detecting Oriented Objects through Group-wise Rotating and Attention

Artificial Visual System for Orientation Detection Based on Hubel–Wiesel Model

Arbitrary-Oriented Scene Text Detection Via Rotation Proposals

Adaptive preprocessing of character recognition image based on neural network

Rotation-Sensitive Regression for Oriented Scene Text Detection

A New Neural Network Oriented Speech Recognition