Related Experiment Video
Updated: Dec 15, 2025
09:59
Fabrication of Flexible Image Sensor Based on Lateral NIPIN Phototransistors
Published on: June 23, 2018
8.1K
Gate-tunable van der Waals heterostructure for reconfigurable neural network vision sensor
Chen-Yu Wang1, Shi-Jun Liang1, Shuang Wang1
1National Laboratory of Solid State Microstructures, School of Physics, Collaborative Innovation Center of Advanced Microstructures, Nanjing University, Nanjing 210093, China.
Science Advances
|July 9, 2020
Summary
This study presents a novel vision sensor using van der Waals (vdW) heterostructures that mimics the human retina for efficient image processing. The reconfigurable sensor can be trained to classify images, paving the way for advanced neural network vision sensors.
Area of Science:
- Neuroscience
- Materials Science
- Electrical Engineering
Background:
- The human retina performs early visual processing using complex neurobiological structures.
- Developing artificial vision sensors that mimic retinal functions is key for efficient image processing.
- Van der Waals (vdW) materials offer unique electronic properties for novel device architectures.
Purpose of the Study:
- To demonstrate a prototype vision sensor emulating retinal functionalities.
- To achieve simultaneous image sensing and processing using reconfigurable pixels.
- To enable image classification directly within the vision sensor.
Main Methods:
- Utilized gate-tunable positive and negative photoresponses of vdW vertical heterostructures.
- Engineered the sensor to emulate bipolar cells and photoreceptors, including their connectivity.
- Implemented pixel-level gate voltage tuning for reconfigurability and training.
Main Results:
- Successfully demonstrated a prototype vision sensor with emulated retinal functionalities.
- Achieved simultaneous image sensing and processing through reconfigurable pixel arrays.
- Showcased the sensor's capability to be trained for image classification by adjusting gate voltages.
Conclusions:
- vdW vertical heterostructures provide a promising platform for developing advanced vision sensors.
- The developed sensor integrates sensing and processing capabilities, mimicking neural networks.
- This approach offers a pathway towards next-generation intelligent vision systems.
